Prompt Engineering Tutorial For Beginners | Learn How To Train AI to Think Like You | Simplilearn

Great. Um, now we're also live on LinkedIn and YouTube and we have more uh participants joining us from uh different countries. Uh it's great to have you all here. Hi, you're a student. Let us know where you're joining us from. Um, hi. Is there someone joining us from Zambia? Great. Hi Sandal, you're joining us from Pune. Hi Omar, you're joining us from Kenya. Hi Amita from BA. Hi Rabi, you're joining us from the US. Good morning to you. Uh, there's someone from Nepal. Sorry I missed the name. Um, okay. We have someone from Nigeria. Hi Ravi, you're joining us from the US. Anupama from Nepal. Hi Srushi from India. Hi Vin California. Great. Wow. Amazing. Amazing. It's it's really great to have such a lively bunch of audience here. Let me quickly introduce myself. Uh I'm Alana and I will be hosting the session on behalf of Simply Learn. Uh we have a super informative uh insight path workshop. Um this is going to be 3 hours long. So please do stay tuned. There's a lot to take away from this session. Um yes to make sure today's session runs smoothly and that everyone gets the most of it we have a few rules that we request everyone to follow. Uh number one please drop any questions that you have in the Q&A box and not in the chat box. We keep getting a lot of messages in the chat box as you can see already. Uh so we might end up missing uh yours. Please use the Q&A box effectively. Um and please do give us some time to get back to you with an answer. We have a lot of participants and viewers here. Um so uh give us some time to get back to you with the quest with the answers. We will uh try our best to address all your questions. Um that said if you have any issues or like you know concerns or doubts uh that you have while following along with our expert you can um you know post that in the chat box as well. Um but uh however avoid sharing any external links or personal details or ill irrelevant um you know messages in the chat box uh please keep all the conversations relevant to the topic only. We will be strictly monitoring this and removing any irrelevant u messages. And uh finally for those who participate in the session till the end and give us your full name in the poll that we'll be launching at the end of the workshop we will provide an attendant certificate. Um you can share this on your social media, add your resume and even update on LinkedIn. Um but please do ensure that you give your full name to us in the poll that we'll be launching. Unfortunately, we will not be able to take any manual submissions, manual entries for the certificate. And along with the certificate you will also be getting a recording of um the um workshop so that you can like you know refer to it at your convenience um and um you know use it to um uh build along as you like. Okay. Um so just curious to know has anyone here attended a simply learn webinar or workshop before or is this your first time? Um let us know in the chat if this is your first time. We're very eager to understand like you know where everyone is uh joining us from. Okay. Someone I see a few responses. You're saying it's first time. First time. First time. Okay. Okay. There's someone here who has attended quite a few workshops. Okay. Shaker that is. Thank you Shaker for joining us again. Um okay third time. N thank you for being here again. Great. So most of you are saying first time so allow me to quickly brief about Simply Learn and what we do. Uh we help people build relevant job ready skills. It is um you know as simple as that. Uh and in this journey we have helped more than 8 million people across 150 plus countries um to level up their careers. Um and we work with some of uh the world's top universities and companies to make that happen. And our idea is very simple. It is to help build skills that are current, practical and actually used in the real world and that's exactly um you know what today's session is also designed uh to reflect. Like I said uh we have an ecosystem of great partnerships. We've built our curriculum in collaboration with highly ranked institutions and global tech giants like Google, IBM, AWS and Microsoft. Um these partnerships really you know sort of uh shape what we teach and how we teach it which means what when you learn with us you are learning what the industry requires and expects right now and um you know the results speak for themselves. Our learners report an average uh 50% salary hike. We maintain an 80% graduate rate which is you know much higher than industry average for online learning and a 4.8 out of five rating from our community. Um this is all because our programs are live hands-on and taught by practitioners and not just academicians. Uh you'll work on real problems not just like you know uh textbook exercises and uh yes so that's about it about simply learn. So we'll get to one of the most exciting parts of the workshop which is uh meeting our experts. It is my pleasure to welcome uh today's uh speaker today's guide uh Kino Beard. Uh Kino is a technical educator and the director of professional learning at A+ virtual learning uh where he designs and delivers courses in Python, data science, machine learning and generative AI for everyone from um K12 educators to working professionals. Uh what I personally love about Kino's approach is that he doesn't you know just teach these concepts, frameworks or tools. He has also built them. He has led multi-month boot camps that take complete beginners um all the way to building real AI applications and he's deeply focused on responsible practical AI usage. Uh so it's going to be a great uh session today. Welcome Kino. It would be great if you could um add on to that introduction and say your hi to our participants. Hi everyone. I want to uh thank you all and welcome all of you here. I'm excited to uh spend um part of my day with you. Uh we're going to be doing getting a little bit uh hands-on with prompt engineering and um looking forward to working with all of you today. Thank you. Okay. So, I want to announce something exciting for all of you here. Um you're getting a free resource after the end of today's session and it is uh ready to use prompts for every job function. Um these are like these just like very generic prompts. Um they are organized by job ro and you know pretty much like ready to copy paste in and each one comes with a note on what it's optimized for and a variation also that you can adapt. Uh so please do stick around till the end uh because this uh alone could save you hours of trial and error with um you know working with AI. Okay. Um before we dive into the how, um I think it's worth spending just a moment on the why because the numbers here are pretty striking. 75% of knowledge workers are already using AI at work today. Um Mckeny estimates that 60 to 70% of work activities can be automated or augmented with AI and Stanford research suggests AI can improve productivity by up to 40% when used effectively. And that last part is the key phrase here when used effectively. U because clearly just having access to an AI tool doesn't automatically make us more productive. Um the way we use it matters uh a lot and that's um that's really the you know entire premise of today's workshop. Uh I'd love to know what um exactly you know brought everyone here. Um what is the number one reason for being here? Please do like you know uh let us know in the poll that we have just launched. Um A is it that you want to get better outputs from AI um or B you want a structured system of working with AI tools or uh C you want to make use of the tools that uh you have to its maximum maximum potential or D do you want to strengthen your JI skills. Um like I said we've got a poll up uh on your screen with four options. Uh please do take a second and let us know which one resonates the most with you. Okay, we have a lot of people voting already more than 100 people have voted. Okay, it's it's pretty split I would say. We have responses um leaning towards getting better outputs from AI but also that you want to develop and strengthen um you know your JI skills. Okay, great. So, it's it's honestly an interesting mix of answers and um all four of these reasons point to the same underlying need. Um which is most of us are using AI but we know we could be getting so much more out of it. Um that gap between what AI can do and what we are actually getting from it. That is what prompt engineering closes and that's exactly what Kino is here to help us with. Okay. So um in the next 1 hour this is how uh sorry next 3 hours this is how like we're going to uh structure our time today. Uh we are going to move through six key areas starting with understanding about what prompt engineering actually is where it is non-negotiable and its impact and use cases across industries and roles. Um then the real meat of the workshop begins. Um we are we've split it into three parts which is foundations advanced frameworks and then expert uh strategies of prompt engineering and Kino will be demonstrating all um of these parts live and yes so let's start from the very beginning um you know I think a lot of people in the audience um you know have heard the term prompt engineering we all uh know about it um it's been thrown around a lot recently but uh I doubt if like you know um all of us have a a very clear idea of what it actually means. So, can you please like break it down for us? Like what is prompt engineering and why is it such a big deal right now? Well, I mean right here. So, if you can just go back, I'd love to read what was on the deck prior and I'll speak from that. So, just go back one um deck. So prompt engineering is the art and science of communicating with AI to get reliable highquality outputs. So um one of the first things that I always tell my students when they are working with AI is garbage in garbage out. Okay? So if you put something if you put garbage into an AI tool you're going to get garbage out. So prompt engineering is really about crafting effective prompts so that you can get what you want from the from the AI from the uh large well the it's a it's a large language model. Um I'm not sure how many of you have heard of the term LLM before but what it is it's a large language model. Um and that that's what it is. So you can't just write anything into it. Okay. um if you put a high quality prompt into your AI tool, whichever one you're using, you're going to get a high quality output as well. So that's the core and essence of prompt engineering. You can move ahead. All right. So when we're looking at prompt engineering, we are thinking about a few things. So we say that prompt engineering really sits at the intersection of um critical thinking, communication design and iterative problem solving. So critical thinking um knowing what you actually need, what are all of the details that I need? How can I be completely thorough when I am prompting my AI but not thorough but not overwhelming? Okay. Uh, communication design structuring how you want to ask it. What's the process flow in which I want to um utilize in terms of how I'm going to ask access thing and once we get into another um getting a little bit more throughout the demo, we'll talk a little bit more about that structuring aspect of it. And then the third aspect is uh iteration. Okay, we're not just doing it one time. We're constantly taking it through a refinement process to making it get uh continuously get better, continuously um continuously improve. So prompt engineering is the practice of designing inputs that guides the AI to produce the most accurate, relevant, and useful outputs possible. So you want to think of it as a difference between asking a question and asking the right question. can move on. Uh Kino I have a question here. Um so if you know I allow me to be a little skeptical. Um you know isn't isn't prompt engineering all about like you know knowing how to type a good question or like a good com command. What is why is this sort of like a skill you know worth learning and not something that you can just figure out naturally over time? Well, I mean, every tool, every AI tool, there there's a way in which you should use that AI tool in order to be um in order for that tool to be most effective. I think that most people I mean in the poll here, if you can just drop in the chat, how many of you are using in your day-today life, how many of you are using AI? Whether that's work or whether that's school, how many of you are utilizing AI? Not not just by watching videos uh of AI slush on YouTube or Tik Tok, but how many of you are utilizing AI on a daily basis? Coming in. Yeah, we're getting Yes. daily. you we're using it every day and we're using it every day to I I say say one in my workflow in my job I use it every day whether it's to write an email to correct my grandma um crafting an email to my colleagues whether it's whether I'm working on a professional development plan for the teachers that I work with whether I am working on sometimes I I somes Sometimes I teach history as well to uh some students whether I'm working with my history students to really organize content. So it's a skill that is absolutely necessary for one I would say one for you to stay competitive in the in the workplace in the the workforce. So here's some stats that we have here. Um prompt engineering directly impacts AI performance. So structured use of AI improves your output uh quality significantly by 40 to 70% in a complex task. Okay. So the better you learn how to prompt, the more effective you are going to um be. Uh it says AI productivity games depends on how well it's used. So AI users complete a task 40% faster and with 18% higher quality and AI benefits vary significantly based on user capability. Okay. So if you are just now hopping on to uh um an AI tool. So for example the first um about two days ago I introduced one of my friends. I actually live in uh Rwanda, Central East Africa and some people here are using AI but not a lot of people are using it and I I introduced chat GPT to one of my good friends and he's never used it. So imagine him going on to chat GPT for the first time and just trying to figure out how it works. His prompt engineering skills are absolutely near zero. But if you are using it over time, he's going to learn how to get better outputs from the AI. So this last stat says here that AI benefits vary significantly based on user capability. If you're new to AI, you really don't know a lot of how do you craft an effective prompt. So AI increase increased productivity by 14% on average but up to 34% for less skilled users. And one of the secrets too about prompt engineering, you can use AI to help you be able to craft some of the prompts that you are um you want. And especially if you are working if you're working in if you have a personal account where maybe you have a subscription or something like that it begins to remember a lot of the things that you have done and it can help you craft and shape those prompts as well. But coming back to your question, uh prompt engineering is a absolute necessary critical skill to help you stay competitive, to help you utilize AI appropriately and for you to get the outputs that you want to be able to solve the problems that you need to solve. Perfect. Thank you for that answer, uh Kenu. And uh moving on. So when we you know we have the word engineering attached to this skill we sort of think that this is relevant for um you know AI engineers or data scientists or ML professionals right so um but prompt engineering has very wide applications and impact across different roles and industries in reality so can you please take us through that so I mean whatever your industry is um there are ways in which you can utilize prompt engineering. So whether that's marketing and communication, so this shows up in terms of uh writing briefs, writing copy, uh you can design an entire marketing uh campaign using the right uh prompt engineering skills. Uh if we take a look at something like HR and uh learning and this is a place that I know a lot in because I tend to I uh work a lot of in HR hiring um folks specifically hiring teachers when I want to write my job descriptions when I want to write and develop my training content even when I'm working with my uh students at SimplyLearn uh I use uh prompt engineering ing um skills when I want to do performance reviews for um my teachers as well uh in whether you're working in let's say finance and operations whether you want to write reports do some sort of analysis uh um even in my personal life my wife runs some logistics operations and I use that to help her um do that kind of logistics and supply operations um for her. So it's not really limited to uh one role. Uh even as a developer um when I'm writing code, how do you get good and proper um proper code and how do you structure it to take a learner from part A to part B to part C? um that is uh another way in which we can utilize pro engineering. So in essence it integrates into everything and um a lot of the things that we are already um already doing. Okay. Thank you Kino. Um I see a lot of repeat questions about the recording. So let me just like you know address that uh uh once again. Yes, this workshop is being recorded and the recording will be shared with everyone on email. All of those uh people who stay till the end and participate in the poll that we will be launching. Uh so please do stay tuned. You will be getting the recording on email. So you can refer to it at your convenience and learn at your pace as well. Um and uh I see a couple of questions about uh attendance certificate and yes uh you will be getting a participation or an attendance certificate for being part of this workshop um as well but please do stay tuned to the end of the workshop. Um and I see a couple of people talking about some uh echo um issues. I um I will I I'm not sure if it's if it's coming from my end because it is only um the um only Kino and I who have our audios on. I will ensure to keep my audio off whenever uh Kino speaking. So um you know to ensure that all of you are hearing um everything clearly. Uh do you have a question I want to address before you move on? Here's a question that I want to um address and this is from a it says a high school student graduating in 2030 and they're worried that prompt engineering might be obsolete as AI gets smarter and it needs less instruction. What is the core underlying skill behind prompt engineering like critical thinking or problem framing that I should master now so that my skills don't expire in a few years? I have a background in competitive debate and essay writing, not computer science. Since prompt engineering is essentially communicating with machines, how can I turn my traditional debate and uh allocation skills into a superpower for writing highly logical expert? Oh, that's a really good question. So, um I would say that the interface itself will change, but I think the underlying thinking skills um won't. So prompt engineering is is really not the skill. I think structured thinking is the skill. So as a as a debater um and a and a writer, I think that you do have some of those structured uh thinking skills as well. But what are some of the things that you should um master? I think you should be able to master things like problem framing and um what I would call um cognitive decomposition. So this means uh being able to define what the real problem is, not just looking at surface questions, being able to break that into some solvable parts and being able to think of how you want to um structure your constraints and your assumptions as well. Um and that your goals are really clearly stated. So AI is getting better at guessing, but that really depends on um your overall one your clarity of intent um the quality and of your structure and your overall precision um of constraint and that's where uh skills that human beings still uh still and I think for a long time would still be able to dominate. So, you're not just like learning to talk to AI, you are learning to to to um really think in um systems. So, and like I said, your debate and essay writing, um, that I think that puts you ahead of most people, you know, uh, it helps you to, uh, understand things like claim, evidence, and your reasoning structure while, um, you're also probably anticipating your counterarguments, and you're able to sort of kind of build a logical flow. This is really what uh high level prompting uh is about. So I hope I kind of addressed that um a little bit for uh for you to high school student. All right. So thank you and thank you Mi. Yeah. For the question. Um Ko before we uh go on to the section I just want to address uh one thing. I see a lot of you pinging your email ids on the chat box. Please don't do that. We will not be able to take any manual entries on the chat box and give you the certificate. We will be launching a poll for this specifically at the end of the workshop. If you give your full name um in that poll then you will be getting the um certificate for the workshop. So please stop uh sending all the uh email ids in the chat. Thank you. Um yes uh Kino like so now that we've set the context we'll move on to part one of the workshop which is foundations of prompt engineering. Um I believe you have eight rules or you know principles on how you you how you can actually build a good prompt. Um over to you uh Ko please walk us through the pointers and let me know when you would like to you know share the screen and uh start the I will share my screen. I just need some actually I can you don't share yet I'll um give me permission but uh you keep sharing I'll just send a request for you to share in the meantime. All right so let's start with some core prompt framework. All right and here we have eight rules and by today we are going to go through about 20 rules. So you should be maybe grabbing some screenshots of this and so on um before we move on. So let's go through the first eight rule that which is some core prompt framework. Uh rule number one we want to be as specific as possible. So the clearer your ask the sharper the output. So, we don't we we don't want to be ambiguous. That's the big idea here. We don't want to be ambiguous. We want to be as specific as possible. That's rule number one. Rule number two, we want to define the role. All right? And what we mean here, we want to be able to give the AI an identity. Okay, it changes the tone, the depth and the perspective instantly. Okay. So if my role is my role is I'm a marketing manager, I am a product develop developer, I am a teacher, I am a student, you want to give the AI some sort of identity because this allows for one tone, depth and perspective. So are we clear on these two rules so far? being specific and being able to give the AI an identity where we are setting the tone and the perspective. Moving on to three. Um in terms of context, we also want to create a context. All right. And this sort of kind of goes goes along with um identity as well. But in that context you can say I am teaching a class this is for a um this is for a this sort of particular business so that the AI begins not necessarily to think for you so that it can find the solution that you are looking for. Number four, you want to be able to state the format. And here we are specifying the structure. So whether or not you want bullets, you want a table, you want steps, you want a summary, you want it to lay out the pros and cons of something. You want to ensure that you are stating the format that you want. You're doing a blog, you are doing uh um a brief, you're writing an executive report for uh seuite personnel. You want to ensure that you are stating the the format. Number five, again, we are setting the tone. What do you want? Do you want a professional tone? Do you want a casual tone? Are you looking to be technical? Are you looking to be very simple? Are you looking for be being persuasive? Um are you looking to be humorous? Okay, so just one word can save you a full rewrite of your um your prompt. So you want to ensure that you are setting the tone. All right. Next, you want to break down the task. All right? Too many ands equals too many task. All right? So, split it. One prompt, one job. All right? So, you don't want the tool, the LLM that you're using, you don't want the LLM to get quote unquote overwhelmed, okay? Where it's doing too many things at a time. So as much as possible, one prompt, one job and you can tear it in that manner. All right? You can show examples. One good examples be three paragraphs of instructions. Okay? So maybe you've done something before on your um in the the AI that you are utilizing and it gave you a good result. Copy and paste that okay so that you can get something similar uh to what you have done because you are showing examples. And next rule, let's use some constraints. Let's be able to set some boundaries in terms of the length, the scope, the vocabulary, and the perspective. Okay? If you want three sentences, ask for three sentences. If you want three paragraphs, ask for three paragraphs. Okay? uh what's the particular scope that you're looking at? What kind of vocabulary that you want it um want it to use? So, sometimes when I'm teaching my K K to 12 uh students um some of them some of my students who might be in quote unquote 11th or 12th grade, sometimes they don't have a the vocabulary of a 11th grade or 12th grader. and their vocabulary might be that of a seventh grader. All right? So I might ask it to um do it at a particular Lexile level where a seventh grader can understand and then you want always want to give it the perspective that you are are looking for maybe based on the audience that you are preparing for. So, eight simple rules, okay, that you can utilize when you are engaged in prompt engineering. One, being specific, defining the role, giving contents, state the format, set the tone, break down the task, show examples, and use constraints. All right. So, which out of these uh eight do you think are which out of these eight do you think would be most useful in your line of um work? Which out of these eight? All right. So, we can move on. Uh Kino, I just wanted to let you know that I have um I have disabled uh the chat option uh every uh because there are a lot of email ids and full names coming in for the certificates which um you know sort of interrupts with the flow of the session. Um so I saw a lot quite a lot of requests as well. So only um you know the host and uh the speaker will be able to see all the live messages coming in the chat box and we have a lot of responses that have come uh you know a lot of them are saying one um actually one is the most uh unanswered uh rule that I can great okay so I'll stop this uh screen share now and you can take it forward from here okay so We're going into our first demo, right? All right. So, let me share. All right. So, I am going to be using Claude. I am I'm actually going to be using I'm I'm a very new user to Claude and give me a moment to just share my screen. All right. So, I'm a new user to Claude and I'm using it for this demo specifically because um I didn't want to use let's say chat GPT because I've been using chat GPT for the past uh 3 years and it knows a lot about me and I don't want it to be influenced by a lot of the things that I know and the things that I do on a daily basis. All right. So, I'm going to be using Claude for this demo. Let me know if you're able to clearly see my You're able to see my screen, right? All right. So, I'm getting some thumbs up that you are seeing about my you're able to see screen. All right. So, we're going to begin with a bad prompt. We're going to begin with a bad prompt. And that prompt is just going to say, "Write something about AI in business." I'm using the free version of Claude here. So, I hope that I don't get any timeouts or anything like that. So, let's just look at a a bad prop. And this is how most people use AI. Very general and very open. And remember, our first rule was being specific. So, let's see what we get when we say to write something about AI in business. Let's see what happens here. All right. So what do we have? So AI where AI is making the biggest dent automation, efficiency, smarter interactions, decision intelligence, hiring and what do you call it? So one it gives me quite a bit of information. Okay. So it's saying the core story about AI in business right now is one of genuine transformation and so on. So the output here is not wrong. Okay. But I would say that this output isn't very useful. Why? Because it's maybe very general. It's too um broad for my specific task. Um, and I really didn't control anything here in terms of what what am I what I'm getting. All right. So, that was an example of a a bad prompt, a poor prompt. Let's work on a better prompt. All right. So I'm going to write here act as a business consultant. So first I've given it a role and now I'm going to ask you to write a five bullet executive summary explaining how AI improve Improves operational efficiency in a small retail business. Use simple language. Let me get my spelling here on inclus and in a small detail in both in small. Let's see what we get here. So, here's my executive summary. AI for operational efficiency in small retail. I'm not going to read it all, but it gives me five bullet points. That's what I asked for. uh smarter inventory management, faster and more personal customer service, less time lost to admin work, sharper pricing decisions, early warning on problems. So the output that I wanted is the output that I got. So as opposed to the first poor prompt where I asked it to write something about AI and business and it g gave me a wide range of things from operations to customer experience analytics to hiring. Uh this one gives me something specific that I want. I want to know about smarter inventory management. AI tracks what's selling and what's sitting on shelves, then automatically reorders the right products at the right time. This cuts down on overstock and ties up cash and stockouts that drives customers away. Less time lost to admin work. Tasks like scheduling staff shifts, processing invoice, and generating sales report can be automated with AI tools. This frees up owners and managers to focus on growth instead of instead of paperwork. So, uh much better prompt uh much clearer output here. I stated the format. I asked for um for for bullets. Okay. I gave the prompt some context. Okay. I defined my um role. Clearly, you don't have to use all eight of the rules at the same time, but it gives you uh something to work um work in. Okay. So, same topic, we get a completely different level of um response. So, um, you can drop in the chat what are some of the things that was changed? What what are some of the things that changed? What do you notice? So, the first prompt was structurally designed to be a poor prompt. I like the second prompt. The second output was concise. Like the first prompt was generic, you know, designed to be a poor prompt, you know, very generic. So, it went all over the place. So, you you have to be the the I would say the track on which the AI train runs on. Okay? And you're going to direct it to where you want it to um to go. All right. So, that was our first demo and we will uh come back with a few we have two more demos to go, but uh I'll turn it back over to you uh so that you can we can I think we want to take a short break at this point. Yes, Kito. Thank you so much. If you can uh um stop the screen share. Stop sharing. I'm going to stop sharing. Yes. Okay. Okay. Great. Um, so, um, maybe Kino, we can look at a couple of questions, uh, that have, uh, you know, come our way. Uh, sure. Do you want to read them out to me because I'm scrolling back in the chat? Yes. Um, okay. So, there was one interesting question on um, you know, how to get like quality AI outputs uh, without without paying. There are a lot of concerns that I see around um you know um losing uh uh tokens very quickly um having uh you know credit limit issues um and having to subscribe uh with a certain amount to get the best quality output. So what are your uh uh insight suggestions on this? Well, what are my insight suggestions? I mean I I mean here's what people pay for what you get what you pay for in a 21st century capitalistic society. We get what we pay for. And I mean there are a lot of tools out there that um people want you to try their tools. So they're definitely going to give you the free demo. Take advantage. use the free demo. But the other day I took my car to the car wash and um it was the cheapest car wash possible because I did my wife just wanted to get the dirt off the car and she paid um here in Rwanda we use Rwanda Franks and she paid 5K to get the car washed and when I when the car came back I was like look at the condition of this car. They didn't wash this car properly. What did they do um to this car? And I was like, she was like, "Hey, you you you paid for what you got." Because a few weeks ago, we took the car to another car wash and we paid 13K and the car and it was well detailed. That's what we paid for. It's the same thing here, you know. So, I can't tell you, well, when the tokens run out, what do you do according to what you do? If you know you need AI to use it to help you do your work, um if you have to pay a $20 a month subscription, if you can afford it, do so. You know, um the other option would be get five different email accounts and when that token is done, jump on to another one. But there there I don't think there's an an easy answer to say, hey, it's a tool. It's a product that folks are using and in order to effectively use their product, uh, you have to pay for the product. Well, good. Okay. No, thank you for that. Um, and one more question before like you know we uh wrap up the first part. Uh, there was a concern about um, you know, uh, the length of the prompt. Usually uh some of them were sharing that um you know in in in the interest of being as detailed as possible uh they end up giving like very lengthy prompts which becomes a little complicated for the AI to understand. So where do you uh stand on like you know a detailed prompt versus a lengthy prompt and I think we'll address that a little bit more um earlier. I would say take it in in in steps. Okay. know what your end your what know what you want your end output to be and then you have to break that down into um different parts because for example my wife runs a a logistics depot beverage business and um I was attempting to generate um like a a what do you a huge dashboard for her so that she knows what's happening at all times across different aspects of the business and um in doing so it was overwhelming the AI. So you had I had to say I had to break that task down into do this piece first after that piece is done do this other piece. So know where know how to chunk your um know how to break it down and then chunk it so that you get the output um that you want because you can overwhelm it quite a bit. Okay, perfect. Thank you, Keno. Um, and yeah, so with this, uh, we are at the end of the part one of the workshop. We can take a quick uh five minute break and uh come back um in exactly five minutes and then we will resume with part two where we're going to cover a lot more of advanced uh prompting uh framework uh advanced prompting rules uh rule u rules as well as uh expert uh prompting strategies. So yes, please do stay tuned and we will resume in just about a few minutes. All right, thank you. Hey guys, uh welcome back everyone. Let us know if we can uh get started with the next session. If you're here just uh um you can react with a thumbs up so uh we know that everyone is set to move on to the next part. Okay. Okay. Great. Uh before I hand it over to Kino just allow me a couple of quick minutes um to talk about something exciting. I have a couple of questions about uh you know asking about uh more sessions if we have like you know uh build sessions that are coming up. So before we go on further um you know we just want to quickly talk about a couple of great sessions. Um the first one is coming from Oxford sets business school. Um it's called data decisions and dig digital transformation. Um if uh today's session has you thinking about how AI fits into you know bigger business decisions and organizational strategy um then this one is uh definitely worth your time. It's being led by faculty from Oxford and it's really about like you know how professionals can position themselves to lead in an AIdriven uh environment and you know it goes a lot beyond just uh learning how to use AI tools. It's uh coming on the 21st of um April. Um it will be at 6:30 p.m. IST. Um and I would share the details of it on the chat right now so that if you are interested, you can scan the QR code or um click on the link and uh register. Yes. And we have another exciting session coming up which will be a demo on how to catch credit card fraud using PowerBI and Genai. Um this has a lot of analytics, a lot of uh um you know um hands-on uh understanding of how um uh fraud works and how you can detect and prevent fraud through um analytics. And this is happening on the 22nd of April. Um, this one is perfect if you want to see AI applied to a data problem in real time. Um, so you will watch an AI engineer and data science instructor Solomon Promise explain how to build a fraud reduction workflow. Um, this is a great way to see how the prompting principles that we are covering today translate into technical applications. Uh, both these sessions are absolutely free to attend and the QR codes like I said are up on your screen right now and the links are in your chat box. I definitely hope to see um all of you or some of you there. Um okay, so moving on uh back to the main stuff. Uh Kino, we can move on to the more advanced territory here. So in this part two of the workshop, Kino is going to demonstrate how to go from basic prompting to precision prompting. Um so what separates someone who is good at prompting from someone who's genuinely expert level kino? I believe you have some rules and principles filled for this segment as well. So, um, please take us through that and then over to you to, um, demonstrate how it works as well. All right. Do you want to just share slide seven? Um, not slide 17, slide number 21 for me. Sorry, I I can you share slide 21 so I can talk about positioning? Okay, just give me a second. Sure. Yes, I am at slide 21 over 20. Yes. All right. If you can just keep it up on the screen uh for me. Do you have it up? Yes. Give me a second. I think it both shared. Just give me a second. I'm not seeing it, but Yeah. Just allow me to Yeah. if if it's a little difficult for you. Let me try doing it instead. Just a second. Apologies for the delay. It should be up. Okay, I got it. I got it. You don't don't worry about it. I got it. All right. So, here let's talk about precision prompting. And we have a couple more rules uh for you. And we're going to go through rules 9 through 15. So, first rule, ask for multiple options. So, never settle for the first output. Requesting alternatives gives you range, comparison, and creative choices. So don't just deal with or take the first thing that the AI gives you. Ask for various options. Say, "Hey, give me three choices to pick from." Maybe you're asking for you're asking for it to design uh some sort of graphic for you. Ask for different um styles. Ask for different options. Rule number 10, use iterative refinement. So your first prompt is a draft. Treat it that way. The best output comes from a conversation, not a single exchange. So you're talking with your LLM like not like you're talking to a person, you're talking to an LLM, but make sure that you are going through the iteration process, okay? where you are refining um your prompt and you're pushing the L LLM to continuously get better. Rule number 11, ask for missing details. So if you're not sure what context to provide, let the AI ask you first. And you can simply ask the AI, hey, what am I missing here? Rule number 12, use stepbystep reasoning. So for complex or analytical task, instruct the AI to show its thinking. And this dramatically improves accuracy. Sometimes when you're utilizing uh let's say I I know a lot about chat GPT you say it's it will show you I'm thinking this thinking that thinking that you really want to see that step or you can ask it hey break down your thought process that allowed you to arrive at this particular results. Rule number 13, avoid ambu ambu ambiguity. Ambu um every vague word in your prompt is a place where AI makes an assumption and it may not be the right one. So you don't want to be ambiguous. Okay? And this goes back to being as specific as possible. Okay? So you don't want to leave the AI to wonder because sometimes when you leave the AI to wonder, AI suffers from something called hallucination. It may do that, okay? Because it wants to give you an answer, but it might not be the answer that you um need. So you don't want to be ambiguous. Number 14, clarify your audience. So the same information that's written for a CEO reads completely different when it's written for a new hire. So you want to define who will read it. And number 15, restrict unwanted behavior. Telling the AI what not to do is just as important as telling it what to do. So one of the things is the AI is not a a real person. Okay. And rule number 15 is very very important. Okay. Because what the AI is doing, it's just utilizing a bunch of different math to get you the response that it wants you to get. It's been trained on everything. There's there's a AI tools out there that's been trained on everything. You know, if you take Wikipedia, there's some AI tool that's being trained on it. So, there's a lot of information in there, but you want the information that you need. So, it's also important to tell the AI what not to um what not to do. So at this point we're going to move over and we are going to demonstrate a few a few more things. All right. So I am going to start a new chat and I want some feedback. not some feedback, but I want some help here. And our goal in this demo is to show refinement and iteration. All right. So, we want a new product. Okay. And if anyone can give me some product ideas and I'll pick from based on what I see in the chat. So, we want to create a marketing message for a new product. Okay. And what should that product be? So, if you can just drop in the chat, what might that product be? Let me get a few suggestions from you. What might that product be? A motorcycle helmet. I like that because we take a lot of motorcycles here in um Kegali. Uh so, I will I'll start with that. Maybe we'll try to do another one as well, but let's try to take a motorcycle helmet. All right. So, create I'm going to start with my prompt. Create a marketing message for a new product. A moto psycho. It was a motorcycle helmet. And let's see what we get here. Uh I am sharing my screen. Are you able to see my screen host? Are you a are you not able to see my screen really? So, let me stop share and try sharing again. Some people are able to see it, some people are not. All right, hold on. Give me a minute. Are you able to see it now? All right. So, you're able to see my screen. All right. So, we asked it to create a new marketing a marketing message for a new product, a motorcycle helmet. So, it says, "I love to tailor this for you. A few quick question. Who is your target audience?" All right. So, I would say um something else. I would say moto cycle drivers. Let's see. Right. So let me see for moto cycle right All right. So, here is my ad, my message. Ride fearless. The road doesn't wait. Neither should you. Introducing the helmet name. Engineered for riders who push limits, not boundaries. Built with militarygrade impact protection and aerodynamic shell that cuts through the wind like a blade and a ventilation system that keeps you locked in and cool when it matters most. This isn't just a helmet, it's your edge. Triple certified safety rating, ultra lightweight carbon filter shell, anti-fog angle, wide angle visor, precision fit. On the road, built for riders, built for victory. All right. And it gives me three different messages for this um app. It gives me one that says it's bold and adrenaline, trustworthy and safe, sleek and premium. But I want to then say create three. And it kind of already did that, but let's see what it comes up with here. Three marketing messages for this moto psycho helmet that targets taxi Let's see what the output there would be. So, you're not going to get the same what do you call it that I get. you're not going to get the same outputs if you're trying the same um prompts for that person who wrote that response. So here we get three different marketing messages. We get one that focuses on durability and value. We get one that focuses on safety and family. And we get one that that's about professional um professional uh professional pride. All right, I want to do uh a new just to kind of elaborate a little bit more on this demo. And this time I'll start over and I'll start simple where I say create a marketing message for a new product. So we have a generic product this time. So it's asking what the product is called. So, we're going to do a a new productivity app. And my target audience is busy professionals. And we don't want this part. I'll skip this. So it's new for busy professionals. All right. So here I get three different uh three different messages, three different marketing messages. Uh first one, your to-do list doesn't have to feel like a weight on your closet. Second one, save two hours a day every day. Third one, the productivity app that respects your intelligence. All right, so each message takes a slightly different angle. Option one, it says it leads with empathy, ideal for social ads or email campaigns where you want an immediate emotional connection. Option two leads with a concrete outcome. Great for landing pages or paid search where people are actively looking for solutions. And option three leans into positioning and identity best for brand awareness or targeting a discerning high achieving audience. All right. And what I want to do now, I just want to say, hey, make one make one humorous, one persuasive, and one direct. Let's see what our outputs here are going to be. All right. All right. And the first one is definitely for me because if you go and check out some of my different windows that are open, I have a lot of tabs that are open. So your 47 open browser tabs call. They are staging and intervention. All right. So that's a humorous one and it's clearly telling me it's humorous. Uh the most successful professionals don't work more, they work with better tools and do more. Stress less. starting 60 seconds. All right. So, I get three different uh one. And I mean like this is really one of the first times I'm using Claude and just from my overall what do you call it with it. Um it's it seems like a very good LLM because a lot of the things it's doing it already. Um, it's doing it already for um it's doing it already um for you. All right. So, the idea is we're moving from vague to having some sort of targeted um from vague answers to having some sort of targeting strategy. And we call this process uh iteration. So your first prompt is not your final prompt. It's just your starting point. And the best when you're using AI, um the best users, they don't write better prompts, they refine um they refine their prompts. So, this was a quick demo on uh precision prompting where we uh went a little bit deeper. Next, we are going to look at Let me know if you're able to see my screen again. How to take it up a notch. How to master working with AI. Okay, so we have rules 16 through 20. Uh rule 16 use rewriting instructions. So AI is an exceptional editor. Give it existing content and specific rewrite instructions for fast controlled trans transformation. Rule number 17, request validation. Ask the AI to check its own work before you do it. It can catch errors, gaps, and inconsistencies faster than a reread. Rule 18. Use multi-step commands. Give the AI a sequence to follow, not just a single task. This produces structured layered outputs. Rule 19. Chain prompts together. Each output becomes the input for the next prompt. This is how you build complex highquality deliverables. And rule number 20, clarify the intent. the why behind your request that helps the AI understand what success looks like and calibrate accordingly. Okay, so there's absolutely no need to begin from scratch. Let's say you've been working uh on a particular project in the AI tool of um the AI tool of your choice. All right, let me get back to my slide here. on the AI tool of um your choice and it has memory. It knows what you have been working on. You can ask for give me a summary of all the prompts that I've been working on for this particular project and how to best structure these prompts. You can ask it to do that and that could be your multi-step um command. Sometimes I ask my AI to do a lot of um calculations and aggregations and sometimes it makes errors. You know, it's a good practice just to ask it to hey can you doublech check the numbers that I'm asking you um you for and then it say oh I made a calculation. So always remember that these tools are they're not 100% perfect. All right? So they they they can make errors. So it's important that you ask that you check them, but you also ask them to uh to check themselves. All right. So for this next demo that I am going to do, I found this article the article talks about recruiters admit to using AI robots for job interviews. Four in five hiring managers say artificial intelligent is used at some point in the process. So I want to just copy some of this article. I'm going to just copy some of it and I'm going to dump it into my AI space here. I'm going go grab a little bit more. And I'm going to dump it in there. And I'm going to grab some more. Not all of it, but And then dump it in there. And one more grab. And let me just get a little bit more. That's about it. All right. So, what I'm going to do, I'm going to just ask it to summarize this article. That's my prompt. summarize this article and I would say 60 words, 60 to 70 All right. So, it summarizes my article. It says AI is rapidly transforming recruitment with 83% of recruiters now using it in hiring. While it helps manage high application volumes, job seekers feel the process has become unfair and impersonal with 63% believing AIE recruitment is less fair than human judgment. Meanwhile, candidates are fighting back by using AI to craft applications. Rising competition with applications for job posting with rising rising competition with applications for job posting up 286% is driving adoption on both sides. So now I'm going to demonstrate chain of prompt chain chain prompt prompting. Okay. So I'm going to say now rewrite this summary for a CEO. I would say give me 150 to 200 words. right. So, you see here I have my rewrite in about 150 to 200 words. And this is specifically tailored for a CEO. Now I can say now turn this into a linked post. So here is my post. It's formatted. It has some hashtags as It acts for there's a sort of kind of engagement strategy on this um this post. Okay. I I just need to add in my name and my title. I can even ask that it adjustes the tone to um match my my voice. And finally I can say now critique this post and improve it. Let's see what happens. So, it's doing some selfcritique of its own. own work. And let's look at the critique. So he says, "Here's my honest critique of the original post followed by the improved version." The opening line, "AI is now on both sides of the hiring table is a cliche that has been used hundreds of times on LinkedIn. It won't stop anyone scrolling." The post also felt somewhat corporate and distant, leaning on statistics without much personality or a clear point of view. It read more like a report than a conversation. The ending question is AI making hiring better or just faster was decent but predictable. Overall, it lacked a strong hook, a distinctive voice, and a concrete takeaway that would make someone stop, read, and share. And then it's telling you what it improved. The hook is now provocative. So our hook now is we built AI into hiring to find better people faster. Instead, we may be driving good candidates away. So the hook is now provocative and specific designed to stop the scroll. The post has a sharper point of view rather than sitting on the fence. The statistics are still there, but it's woven in more naturally. The ending is more personal and challen challenging, inviting genuine debate rather than a generic yes no. And the tone feels a little bit more uh human. So it says he changes at a glance. The opening is now a bold specific claim than a neutral observation. The block quote format breaks up the text visually and draws the eye to the most striking stat. The phrase that's not an efficiency gain, that's a trust crisis. Give the gives the post a memorable sharable line. And the closing question is directed specifically at decision makers, making it more likely to prompt a meaningful response in the comments. So what happened here? All right. So we didn't just get an answer. We got more like a workflow. And this is really the difference between beginner and advanced users. Um beginners are using maybe one prompt. your expert users are building systems where they have a workflow that can be repeated over over time. So, always ask the AI um what it needs. Okay? before answering, make sure that you have some clarifying questions to help improve the uh results. All right? And this is probably one of the most powerful things that you can do. Okay? So, sometimes you don't always need better prompts. uh you just need sometimes to be able to ask better questions from the um from the AI. So these 20 rules, they aren't just separate ideas. They are one uh system that should be used um together. All right. Um so here are a few takeaways for you. Uh I think most of us here know know that AI is not magic. What's behind this is a lot of um math um a lot of vectors matrix factorization and things in order to to really build a a good strong large language model. So AI isn't magical. It's a it's a tool. And the difference between getting average results and powerful results is when you know how to use it um well. So some of these skills I want you to take it and apply it to your actual work whether it's your reports uh your data analysis your business decisions uh whatever it is that um that you do um and this is where it becomes a real real valuable skill. So, that's the end of our demo and uh Alicia uh and Anya, I'll turn it back over to you, but I'll still be here. Let me stop share. Yes, perfect. Thank you, Pino. Just give me a second to start the screen share again. Okay. Yes. So these were the umert prompting strategies that we just covered. Um before we go on to the next segment um I got a couple of questions about uh um concerns about not being able to register for um our upcoming um u fraud analytics webinar. So I'm going to share another link here um in the chat box with everyone. should be um you know they should not have any issues. Um so please do register for the session if you're interested to attend and if you have registered it will take a while you will get a confirmation email from uh Zoom um as well as us. Okay. Um okay so we have the um you know Q&A segment coming up uh very very shortly. Um we have a lot of questions interesting questions that have come our way and we will be uh covering as much as possible in the Q&A segment with Kino um very shortly. Uh but before that we would like to um you know just uh um talk to you a little bit about one of our programs. Um so we started this workshop by you know understanding that AI is everywhere but the quality of what you get depends entirely on how you engage with it. Uh we learned about prompt engineering and how it sits at the intersection of critical thinking, communication design and iteration. um it is a skill and um you know it has to be honed and built upon. Uh we also built up from core principles like specific specificity, role definition and format um all the way to advanced techniques like chain prompting, multi-step commands and um asking AI to validate its work. Um and I think the big uh takeaway um is this the gap between someone who uses AI and someone who works with AI as a productivity multiplier. um that gap is prompt engineering. um now today in this workshop we have given you um you know a good foundation of how the prompt engineering works um but if you're thinking that you want to take this further you want to go from knowing these principles to actually building with AI working with LLMs understanding geni architecture that's exactly where our apply generative AI specialization um comes in. So let's uh talk a little bit more about that. Um it is a live online and interactive program that spans for 16 weeks u requiring just about 8 to 10 hours of commitment per week. Um it has a dynamic AI curriculum. Uh you will learn agentic AI deep learning um NLM track pipelines AI uh literacy and a lot more. Um you will get to work on 60 plus excises and seven plus projects with guidance from experts. Um so you walk away with a lot of practical learning such as from this workshop um and you have opportunities to apply that learning as well. Um so this workshop is just like a very very small uh preview and or a demo into what um more the extent of learning opportunities there are with uh this program. Um Kino uh you teach quite a lot of jai programs with uh simply learn. uh it will be great if you can uh come in here and you know um share with our participants how our curriculum is and how industry aligned um our learning path is uh you and how it supports learners. All right. Thank you. So I am an instructor for a number of courses uh here at SimplyLearn um specifically all related to um to AI. And uh one of the first questions that I get and one of the first questions that a lot of um students ask me is do I need to learn how to code? What kind of coding do I know? The program typically takes you from zero to I would say almost zero. But you even if you have no background, no technical background, a program like this is designed for you. Uh one of the things that I like about the um program as the instructor is the the overall design of the program. Um it's really built on some key key learning principles. Um and specifically as in coming from an education background that really works for a lot of learners. There are opportunities for many different at bats at it. So for example a lot of the lectures are already recorded. So in many cases you will get uh you can preview the content before coming to class. Um during class there are opportunities for live interaction with an instructor and at the end of that you can always go back and watch the recording of what you missed. So maybe the instructor was going a little bit fast and you didn't catch something. You can always go back and then learn at your own um pace. In um this specific program, some of the things that you are going to be introduced to includes um you're going to um have a program induction and uh some overall AI literacy skills. You're going to look at some advanced generative AI models and architecture. So being able to understand things like maybe uh transformer model being able to understand the architecture behind transformers self attention and uh positional encoding I like to tell students as well that a lot of people are using AI okay but a program like this makes you know exactly how um AI works. It's like I'm not a mechanic. I just drive my car. Okay. When my car has an engine problem, I take it to the mechanic and the mechanic fixes the car. Okay? Ask me how the car works. I really don't know how the car works. I just want the car to be I just know, hey, you put gas in the car, you press the brakes, and you drive. Okay? A program like this allows you to become the mechanic. Okay? where you're opening the hood of the car and you're getting you're understanding the functionalities of the car. you're going to be able to understand, hey, what neural networks are, how do you train um neural networks, how do you fine-tune things, how do you customize your LLMs, what's the importance of uh data collection, data cleaning, and pre-processing in all of these um things. How do you um deploy um things? How do you um use how do you build things like a a chatbot? How do you how is how is how exactly is Claude or chat GPT generating context? Why is it able to summarize something? Why is it able to perform something like sentiment analysis? So a lot of the the the core of these programs really allow you to become that functional not just a consumer of AI but someone who understands how how how it is working. You know what? We may not fully understand what's happening inside of the blackbox with all of the math and calculus and so on that's involved, but you're going to get a much better perspective than I would say 95% of the population who is just using AI for the sake of hey, it makes my task a little bit easier. Um, you're going to learn things about uh AI governance. you know, there are a lot of ethics uh that surround AI and that's important for for for you to understand so that you're not crossing those uh ethical um those ethical boundaries. You're going to look at things like image gener gener generation c capabilities where you are going to be uh doing projects um like using convolutional neural networks to see how it's going to how it's how it how it's actually able to generate that lifelike um image. So, um, a program like this really I think it's very effective as, uh, an instructor. Um, a lot of my students, they come in with absolute no knowledge and the first few nights sometimes it's a little bit rough, but I'm usually there and our instructors are there to support them every step of the way. We have cohort managers that are there to support learners. um as well, but it gives you a thorough more than an introduction to um using and working with AI. And then it also helps you prepare for certifications like the Microsoft Azure AI fundamentals. And it also um takes you gives you a good thorough um understanding of and using of Microsoft C-Pilot. So, um I'll turn it back over to you, but that's a pretty um what the curriculum uh is like. I mean, there's a lot more that I could have said, but that's just a general framework um of our of our curriculum. It's technical, but it's also doable. Thank you, Kino. I think that was a pretty comprehensive uh overview of the learning path. And I will quickly um you know move on to um the projects uh in in this uh uh program. As we mentioned this program has a lot of practical learning opportunities such as um you know uh building an image generation application, a marketing design studio, an HR assistant and a lot more. Um and what you will actually be learning across this program is genuinely impressive in terms of um you know the breadth of skills. We're talking about everything from LLM architecture and fine-tuning to retrieval augmented generation, agenti, prompt engineering, um jai governance, the full stack of what's relevant um in the market right now. Um Kino we've spent um you know today the last couple of hours talking about prompting but um this slide shows a much broader um you know list of skills and tools. So how important in your perspective, how important are these skills and tools for serious AI practitioners, for people who want to build a career or enter into the AI market? Uh from the lens of a recruiter or an employer, how important are these? Um I would say they're very very important and I think going back to my uh idea of not just using the tool but understanding how the tool works and the these are the things that allows the um the tools to work. So one um the Python programming language is definitely an important programming language in regards to uh using and coding with AI. Very very important. Uh you a lot of our programs introduce you to Python. A lot of our programs the anchor coding language in the programs um is um is Python. Uh and then a lot of the other things are a a a lot surrounding the nature of the architecture of what you are doing things that are going to help you to um deal with uh developing and building your models whether you're pre-training them, you're fine-tuning them, maybe you're doing some aspect of um transfer learning to some of these very um specific application. being able to choose the right model um a pro a program like this is going to help you and all of these topics um together you're going to gain the knowledge and the expertise to um to do that and I and in the previous slide you also showed a little bit about your the different projects and portfolios that's important for employers as um well what are you able to do what are you able to um build and produce. What can you demonstrate that you have done? So, a lot of these tools uh are tools that people need to know how they work. Uh employers are looking at people who not not only can just use the tool but can really build something authentic for a particular business for uh maybe a particular application utilizing um utilizing these tools. Perfect. Thank you so much uh Kino. And um the program is structured uh to work around a professional's life. Um you get live uh weekly sessions with experts for that accountability and depth. Uh self-paced materials for flexibility and hands-on labs where you'll actually um you know be building things with mentor support. Um in addition to this, you also have a pure community of professionals navigating the same journey. Um so this this entire learning environment is designed to keeping in mind your um you know professional life as well. Um and beyond the curriculum there is a whole um you know career support layer built in uh technical assessments to sharpen your proficiency, mock interview prep so you can um you know articulate your skills more confidently, AI powered tools to optimize your resume and LinkedIn visibility as well as curated job opportunities matched to your profile. So you will not be you know just earning a certificate you um you know you will also be getting the environment and the support to actually make use of it. Um so that end to end support is what makes this a good uh career investment and uh so for those of you who are interested here are the program investment details. Uh for Indian learners the fee is $119,9.99. Um and for the US learners um it is uh $2,699. Uh we have installments starting at just 5,373 rupees a month for um learners from the uh from India and for learners from the US the installments uh start at just $270 a month. Um for our learners from other parts of the world, we know that um you know we have participants from all over the world here. Uh apologies for not being able to share um the uh investment details exactly in your currency. Um but you can um use this uh link that I will be sharing in the chat um right now um to get the details of the program and the investment details in your own currency. So with that, I'm going to launch a poll to take your interest in enrolling in the program. All you have to do is vote yes or no in this poll. Uh please do participate and drop in your answers now. Um even if you have any queries, you need more details, you can click on yes. Our expert advisor will be getting in touch with you and guiding you forward. Um and in case if you're voting no, please uh do feel free to let us know why um in the chat. We would genuinely like to know and understand what's holding you back. Um great. Okay. So, uh while the pool is live, um Kino, I'm going to start reading out the questions and we can take a look at um you know, some of the most interesting questions that have come our way. Okay. Um this is one question that I found pretty interesting. Um if AI can help us write prompts, why should we master prompt engineering? If AI can help you write prompts, why should you master prompt engineering? Because um it's the thinking skills that you are getting in terms of being able to craft an effective prompt. It's that I would say it's that interplay between um man and machine. Um I think AI is going to be able to do more in the future but I think um as in terms of as thinking feeling human beings um you have to be able to um do that because AI is not uh unless you expose it to um your the details of your particular situation um what you're going to get is something very if you don't expose it to the details of your particular situation. What you're going to get is something very um generic, but I think it's a very good skill and I think I had addressed that at the very um beginning um as well. It's a critical thinking skill that you as a human um need and um you you have to be the one that is in control of the machine. I would say, you know, a few months ago, I did a podcast with a um with a colleague and we were we're we're doing this series on on AI and the she was very concerned about folks losing their jobs with AI. The people who will lose their jobs with AI is who is allowing AI to do their jobs for them. Okay? But those who can bring solutions to the table with the help of AI and using AI appropriately are the people who are going to survive whatever we think um might be on the verge of of of happening. So it's not allowing the tool to replace you but being able to master the tool so that you're working uh side by side um with it. So it's not just about mastering prompt engineering. It's about that critical engagement with the tool to be able to produce the solutions that you want. That's my take on that question. Perfect. Thank you. Um the next question is from Nirj. Um his uh his point is uh I have heard that different models give better outputs with different structured inputs. Um, is it correct? Uh, yeah, that's the correct answer. Okay, so models are only getting better and better over over time. Some people find their tool and they like their tool and they develop a long-term uh I I lack of a better word, they develop a long-term relationship um with that that tool. Um, and new tools are coming out every day and the different models are doing uh different things. Like one of the other tools that I like to use, I like to use Google Notebook Ellen. I think it's a very very um powerful tool. It's a tool that's not it doesn't hallucinate um as much as let's say like some earlier models of chat GPT where it wants to give you an answer. But the the question there that there are different tools for different things. If you want to be able to write songs using AI, um you you can use chat GPT to write a song, but it's not going to sing that song for you. You're going to have to use another to tool in order to um in order to do that. But that's a very good accurate uh uh good correct line of thinking in regards to the question. Okay, thank you Kino. I hope that answered your question. Um, okay. So, actually we've gotten quite a lot of responses to enroll in the program. So, just allow me a second to um stop the poll. Okay, great. I've uh stopped the poll right now. And yes, we will continue to take in a few more questions. Um and yes, if um for those of you who would like um attendance certificate for the workshop, please do stay tuned. We will be launching um another poll for that very quickly. So give me a second. yes, I'm going to read out the next question. Uh this is from someone who says that they have foundational knowledge in tools like PowerBI but lack advanced coding or DAX skills. Uh what is the best prompt framework to use when asking an AI to help write formulas or build data models that they don't fully understand yet. Um they know a little bit of coding only um and want um you know to further work on their engineering skills. So what is the best the start of the question. Can you repeat the beginning of the question again? Yeah. So, they have a foundational advanced coding skills. Uh, what is the best prompt framework to use when asking an AI to help write formulas or build data models that they don't fully understand yet? All right. Um I would well what I would say is if you have a data set let's say you're you have a excel file or a CSV file I think the coding is not necessarily relevant. Okay. But you can take your data set, dump it into your uh LLM, whatever you're using, whether it's Claude, whether it's Gemini, whether it's chat GPT. And from there, I would say you can begin to ask the AI to begin to prompt you to say, "What's the best way to analyze this data?" That's where I would start from. Okay. I don't think you need to quote unquote learn any quot of any new formulas. I think the AI is going to be able to um do that for you. But it's just I think it's just a data dump into the the LLM and ask it to do the analysis for you. Maybe you're looking for a perspective. Share that perspective with the LLM and it should be able to do that for you. Perfect. Um, moving on to the next question. Considering the effort required to craft effective prompts, um is it sometimes more efficient to complete certain tasks manually rather than using AI? Consider the effort to craft effective prompts. Um so how do I answer that question? I would say that what you're looking for is an output. what you're looking for is a solution to a problem. The the LLM is going to be able to solve that problem way faster, give you a variety of solutions. Um, give you more solutions to your problem than attempting to do it manually. That's what I'm kind of getting from um the question. I mean, you're not you're not going to be able to beat the math behind the machine. That's what I am going to say. You're not going to be able to do that. Um, so I would rather put the effort into guiding the LLM to help me solve the problem and give me a variety um of answers. So, I think it's working um smarter than attempting to do that hard manual work and you're not going to get um oneif of the output that you can possibly get. That's how I would I would answer um that question. All right, I'm seeing something in the chat. I'm seeing something in the chat here. They're saying while using chat GPT or cord or Gemini, it gives me an error like server busy high traffic retry after a few minutes. Is it a common phenomena or can it be avoided? Can you please advise? Um, so are you using a free model or are you using a paid model? That's what I would um say there. It depends on uh the tokens when you're using you're using a paid model. All right. So um so I don't know. I I'll have to look into that a little bit more. I've never really experienced that. I know when I was using the free version of chat GPT, my tokens would um run out, but I've never really gotten that message. I can take a deeper look into that and see, but go ahead. Yes. Uh S, we will see if we can get back to you um you know specifically on that question, on that concern over email. Um unfortunately we cannot uh take this up uh right now but we will try to get back to you on email. Um Kino I will move on to the next question. Um so these AI softwares and AI tools are being trained on our data. So what AI tools should or could be used we can you know installed in our systems and um you know it will give us more security and help us understand and use AI much better. So I think privacy and security is an important uh concern when it's being used. You have to be mindful the data that you share um with them is the is it's your data um that you share but that data is not going to be shared with um others. what it's going to be what it's doing it's helping to make the next model a little bit more efficient, make the next model a little bit um better, but what you put out there is what the AI is going to know um about you, you know. So, um I'm sure I think all of these uh different companies have um their privacy policy, their data policy and everything um like that. And you just have to know the type of user um that you are, you know. Um, that's the best way that I can answer um that question. If it's information that you don't want shared, don't put it into the AI model. Okay? But my AI model is not going to take my personal information and then go share it with somebody else. Okay? That's just between you and your account on that particular platform. Perfect. Um, moving on. Just give me a second. Uh, can it can we use AI tools for academic research um in MSE or PhD program um and not have AI plagiarism test issues? Um, that's a very good question. That's a that's a question on the ethical aspect uh of things and my frame of reference for that question um is coming from I was I I at one point I was enrolled in a PhD program and I didn't did not finish it because I fell in love with data science and I decided to to go down um that path. I am not against anyone getting a PhD or anything like that. There are there are rules to the academy. Okay. But on the flip side of things, um the academy in and of itself has always kind of um they have been they haven't been as sharing with knowledge as they should have. That's my opinion. Okay. Um, someone would complete a PhD and between them and their…

Transcript truncated. Watch the full video for the complete content.