Build a Strong AI Portfolio with the Generative AI Specialization

M right and towards the end we're going to be talking about the career support and job ready outcomes once you complete the course and we are going to wrap the session with a Q&A uh session as well. Uh so you can save your questions towards the end or post them in the Q&A box and we will definitely take these up. Um and a few ground rules before we uh begin. Uh hello everyone for people who've again introduced yourselves. Um, if you have any questions, please put them in the Q&A box. The only reason being we do not want to miss out on any of your questions when you post them in the chat. Due to the number of messages, we kind of miss out on that, right? The second point is please do not share any external links in the chat box. Um, don't share any WhatsApp groups. Keep it relevant to the topic and uh try not to engage in the chat while uh you know we are explaining the speaker is explaining the topic. If you have questions, please put them in the Q&A box. And one of the most asked questions about the certificate of participation, uh you will be receiving it for that. Uh we will be sharing a form towards the end of the webinar. You can fill up your full name and you will receive the certificates within 48 hours. Right? And I hope all of you are aligned with that. Uh let's try to focus on the webinar. Please do not use your mobile phones. It's going to be a short and a really insightful session. Right? So uh out of that uh before beginning I'll just take five minutes of your time how many of you have attended webinars with us before. Uh if it's your first time you can comment first in the chat. Uh I'd just like to see before I introduce uh simply learn. Okay so many people joining us for the first time. Really interesting and really nice to see all of you join us. A few people who uh joined us before as well. For anyone who's joining us for the second time or if you've attended webinars with us, welcome once again. And for anyone who's attending it for the very first time, I'll quickly introduce simply learn. Uh so what we do is we've uh we actually provide uh courses and we have helped 8 million plus learners across 150 plus countries to upskill themselves. And how do we do this? Through global partnerships with universities and also industry partners. Here are some program highlights for all of you. So, uh, most of our learners have reported a 50% salary hike on completing courses with us and we've also been rated 4.8 out of five by our alumni members. So, it also speaks about their learning experience with us. We also have a graduation rate of 80%. So, these are just some of the things that we wanted to uh let you know before we started the webinar. And like you can see on the screen, these are our industry partners and also university partners. Uh so there's Google, Microsoft, T Michigan University. And today's webinar uh we are also going to be talking about generative AI specialization by IIT Madras Praverak. And we'll talk about that more towards the end of the webinar. Right. I hope all of you are still with me. Okay. Um awesome. Uh so now uh this is one of the most interesting parts of the webinar for me. I love introducing our speakers. Uh so we know that uh generative AI is transforming and we really need someone who has a lot of knowledge to speak on this. Right. So uh this is Basel Dakla and I'm really happy to introduce him. He currently leads three high impact analytics teams marketing and growth analytics reporting and visualization and also data quality and automation. which basically means that he works at the intersection of data automation and also real business decisions and it's exactly the kind of uh you know requirements that is required in a speaker uh to speak on the Genai portfolio that actually impresses employers he's also been hiring people for teams worked with a lot of learners across these all these years and that's also one of the most interesting parts that he's a trainer and a content creator and has also collaborated with institutions like UCST and Purdue University uh teaching generative AI as well. So he's also one of our uh generative AI trainers. So I'm really happy uh to welcome Basel to today's webinar. So Basel, I'm sure everyone's looking forward to hearing more from you. It would be great if you could introduce yourself. Just a second. I am just quickly checking with Basil. I don't think he is audible. Just give me a minute. Okay. So, he is just having a few issues with his audio. He will quickly join us. In the meantime, I will be launching a poll uh to understand your experience. How many years of experience uh does all of you have? Are you all uh freshers or uh do you have a lot of work experience? Okay, I can see a lot of freshers and also people with a lot of experience. I'm also connecting with Bassin in the meantime. Just give me a moment. Okay. Sorry about the confusion. He's joining us. Uh so I can see that there are a lot of freshers. Um so many people with up to three years of experience and also seven plus years of experience. Wow that's a great audience. Thank you all for your responses. I am also quickly going to end the poll here and Basel is here with us. So Basel uh it would be great if you could introduce yourself. I've already introduced you to the audience. Thank you very much. I actually got to hear the bit about the uh introduction. So uh I definitely appreciate the kind words and the warm introduction. Hello everyone. My name is Bassel Dlala and uh uh as you saw basically been involved in uh industry of analytics and data engineering for a long time now and uh that was spanned different applications. Uh think about this providing end-to-end solutions from the data infrastructure realm all the way to the data analytics and client-f facing applications uh and advanced data analytics including also providing consultation for various industries and businesses in machine learning and AI implementation and along with that also have experience in training. So it's a passion of mine is training and uh seeing people starting from the basic knowledge of analytics or not even any uh background with analytics and seeing them uh being excel in AI and building machine learning. So it's been a really exciting and rewarding career so far and happy to be on here. Thank you so much for that wonderful introduction and I'm sure everyone's waiting to uh learn more about generative AI and how to build a portfolio. So before we get into those details um I would first like your thoughts on the generative AI boom that's happening. We all know that generative AI is everywhere in every uh stream right now. So what are your thoughts on generative AI uh the kind of growth one can expect in this field? Yeah, absolutely. So I I think we're in a market right now that is very uh competitive and things are shifting very fast. So it's nice to start with just understanding the scale of what's happening right now around us and what is the vision for AI and its implementation in the future. We're in an AI boom and uh I think some people still think that there's an AI bubble. I I tend to disagree just because we're still not there yet in terms of full implementation of AI and therefore we still have a lot of uh excitement and a lot of uh investment opportunities to build AI and a lot of emerging businesses or let's say even countries that are interested in in being a part of the AI boom. Uh so just to start off with a big figure here $968 billion dollars it's almost a trillion dollars by 2032 that is going to be the global market size uh projection for now and this number keeps changing in fact it keeps increasing every time we revisit that statistic. Uh we've got also up to 4.4 4 trillion a year. That is potential annual economic value ad. Uh think about this in terms of how AI is right now implemented in almost every industry. So uh a lot of value ad there. And with that, it's a nice segue to talk about the talent gap that we have. 75% talent gap. This is actually one of the main barriers why uh AI is not accelerating as it's supposed to. because we have over threearters of companies still can't find the right talent for generative AI and other AI and ML applications. They need skilled talent. They know what to do. They uh are familiar with different technologies and able to navigate through the complexity of AI, right? Um and that clearly uh gives us an idea about how massive the generative AI boom is and how quickly the uh skill talent is outpacing the supply like you clearly mentioned. Uh so it also um brings us to a very important point that everyone is learning right now AI right now and they're upskilling in one way or the other. uh but uh that is also where the an EI portfolio makes all the difference right having the right kind of experience or projects it's very crucial in today's time so Basel it would be great if you could share your thoughts on this because you've hired members for the team over these years you've worked with people who are transitioning uh so how why does an AI portfolio matter today yeah uh if you think about it back to the point that we mentioned this AI boom got everyone excited so everyone wants to get on board with it and they want to be involved. So, we're in a market basically uh where everyone claims the same skills. So there's a lot of competition still high demand but still a lot of competition especially with the ease of access and ease of way to to it be it became very easy to claim certain technologies you know for individuals say I have experience with this and that. Uh but this is where building a portfolio is going to make you stand out right portfolios will let recruiters see how you think how you worked on a specific project they can cut through let's say the buzzwords of instead of claiming that I know about this technology and pietorch for example or I've I've done some uh timeline forecasting with online portfolios you can showcase your work Right? You show your decision making because it's it's not only about the technical skills. It's also about your personal skills, your business skills. How do you handle decision making, right? So, it's not just your technical skill. You want to also showcase how you were able to document your work, how you were able to organize it, right? So, think of it as your AI careers live demo to someone. You're trying to demo a product and you are the product, right? You are basically showcasing the great work you've done throughout this portfolio. So high level skills change fast. So you need to showcase how you're able to adapt and work on these new skills and improve them. You give a proof that you are working on these projects instead of just talking verbally about them. And then it shows your real problem solving skills. It helps you stand out in general. And then it reduces the need for experience. Think about this. This is a question I get asked all the time about people like, "Hey, I don't have much experience or background. I've never worked as an AI engineer or machine learning practitioner. How am I going to get a job?" Your best bet is to actually build a portfolio to showcase your experience in AI and machine learning and therefore you are proving that you are able to solve these complex problems. Right? And I think that gives us a very clear idea as to why AI portfolios matter today. And uh that also brings me to the next question. Right now after hearing this I'm sure a lot of them are a lot of the participants are wondering what really makes a strong generative AI portfolio and how one can work towards creating a portfolio that stands out. So it would be great if you could walk us through this and break down what these different elements uh could be when they create a portfolio and uh how they can showcase this to their recruiter or hiring manager. Yeah. I mean also be in the shoes of the hiring manager, right? I've interviewed individuals and I've it's always been a situation where the people who can showcase their work and they can speak very intelligently about their work. It's not just you know copy paste a project and whatnot. But just just imagine for a second you're a hiring manager and you have two applicants. Both list skills like Python uh lang chain some of the technologies are being used in generative AI. uh but only one of them show um shows the work a working rag system let's say or a chatbot or a project they've worked on uh where it actually served a purpose and especially if that purpose aligns well with the industry I mean as a hiring manager you're going to lean towards the person that's showcasing the work right not just seeing words on a resume so um a strong project isn't about just showing and off the code. It's also about showing your uh thought process, your planning, showing how you're able to be uh nimble with technology and coming up with new ideas and solutions because a lot of people right now they do a lot of copy paste. You can see a lot of projects online and people will repeat the same work. So there's no benefit there, right? So recruiters want to see how you design these systems. How do you handle things like safety if needed or if applicable? Uh think through different tradeoffs because there are multiple options when you build a machine learning system or a generic system. Uh and then how you're able to communicate your work. That's very important. That's especially during the interview, right? So they're hiring your work and your thinking and your decision-m capabilities. So we can see here if you want to build a portfolio, you want to showcase the problem framing, you want to showcase your workflow design, your understanding of different LLM systems. And LLM for the ones that are not familiar stands for large language models. These are basically the categories for um it's a category for the chat GPT system, Gemini, all these uh engines that you hear about and use on a daily basis and uh safety thinking. So making sure that you are implementing guard rails for your system. uh this just helps the recruiter understand that you do have awareness of the organization and the governance that comes with it. And then finally, as we mentioned, decision reasoning and decision- making, right? Um and uh those are some really important points that we covered here. uh but uh there's one more thing that not everyone's portfolio so I'm pretty sure a lot of them are already currently uh working on AI or learning new AI concepts but sometimes even if you work on projects not everyone's portfolio communicates those strengths clearly and maybe they are not able to convince the hiring managers about their work like you said communication and having clarity on the work is also really important uh so uh Basel because you have also worked uh I mean uh hired a lot of people in field. Uh could you help us understand uh the difference between a good portfolio and a weak portfolio and what kind of differentiates uh these kinds of portfolios and how one can probably improve their current portfolio to meet the market demands and highlight the kind of skills that recruiters are looking forward to have. Yeah. Um so having a portfolio period is nice but again we keep talking about uh how competitive things are getting and how a lot of people you know are trying to do claim certain skills and trying to do copy paste. So this is where it's nice to um have a goal of achieving an outstanding portfolio. Right? So let's talk about the stages between weak versus good versus outstanding. So what dictates these different portfolios? Well, when you have a weak portfolio, you're basically just copy pasting, collecting different uh scattered segments of an example or a project and just looks like a tutorial style trivial uh type of uh project, right? Uh there's lack of documentation. So, that shows weakness. That means you don't know much about what's going on in some of these uh components and you're not able to explain it well. Uh messy, incomplete code. That's a big one, right? You don't want to showcase a project you're proud of, but still has messy code and code that may uh cause errors because, you know, recruiters may be able to grab your code and try to test it, see if it if it works. and then um no thoughtful steps meaning when you look at the machine learning or AI project pipeline there are multiple steps that you have to consider. So we start from collecting data we start then we do data cleansing all the way to deploying the model then model evaluation. So just deploying the model is not enough. You can't just say, "Okay, I deployed the model. I'm good to go." You have to evaluate your model. You need to make sure that your model is performing well. Otherwise, your company, if they're going to rely on you and rely on your work, and your model is not performing up to the standards, it's just pointless. It's a waste of time. Um, and then to wrap it up, you need to have real world context as much as possible. I mean I see something out online a lot is being used to claim you know you have experience in machine learning is using the Titanic data set. Titanic data set is like the hello world of uh machine learning and you see everyone using that. So you need to come up with a a real world scenario or something that is unique so you can stand out. So that's yeah yeah go ahead. Uh no those were some really uh insightful points. I just wanted to check if you had any other tips for all the participants here uh to improve their portfolios. Anything else that you'd like to add? Yeah, I think uh most people because most people stop at good, they don't think about what is outstanding. So, if you want to have something that is outstanding and what I call unforgettable, you know, someone who's either a recruiter or uh a hiring manager that knows about the tech and all the components, they want to read your portfolio. Make sure this should be always your compass is that make sure that it it is impressive and it's unforgettable. You have a unique solution overall. you are uh adding your touch to it, right? Not just something that is very generic and high level, right? Um thank you for sharing that. Those were some really good points. And I also wanted to u uh kind of interrupt and say that I see a lot of uh chats here. Please do not uh share anything that is not relevant to the webinar in the chat because uh if it's uh external links, we will have to remove remove you from the webinar. and please uh refrain from uh you know sharing these reactions as well. Um so okay coming back to the webinar I think the most important aspect of today's webinar is on how one can build a standout generative AI portfolio and that is something that we are going to be talking about. So Basel what are your tips? Uh so we covered what a weak versus good versus outstanding port portfolio looks like and you also gave us some insights on how one can uh work towards creating an impressive portfolio. Uh so for someone who is a complete beginner and do not have any experience in this field, how can one build a standout generative portfolio? What are your thoughts on this? Yeah. Uh so the first one you can see here it's to me one of the most important uh points so you can move things to the next level is to focus on real world impact right projects in most cases they tell a story and they solve for a problem right so you start from a problem you define the problem you develop an action and then you have results to showcase right And honestly, even if you if you haven't developed a finalized solution yet, that's still kind of okay as long as you explain your next steps. What are you missing? Basically, it just shows that you are aware of what's missing, how can you take things to the next level, it's just you haven't gotten to that step yet, which is fine in in in uh some scenarios. So, that's very important. Number two is use the right tools and techniques, right? Don't try to chase the shiny objects. Uh oh, I've heard about this LLM or I've heard about this uh machine learning engine and I want to apply it directly. Uh because it's just that people say it's the greatest. No, have a thorough analysis and research. Understand the uh pros and cons of that tool. Understand if it's a good fit to your data. And a lot of people ask me during my lectures what is the best machine learning algorithm and you can't answer that. The reason why because every data has its own story. Every data has its own uh scenario or complexity. So you have to always think about a solution that is tailored for that data set and problem. Okay. So would be uh things like understanding different systems like lang chain rag um hug and face platforms right so you want to showcase if you're focusing on generative AI projects you want to showcase how you were able to um build the infrastructure and how you thought about the prompt design the m model fine-tuning if you've applied any fine-tuning and then um you know showcase any solution that goes beyond the API uh system and uh number three is building an endto-end project. So if you remember when I mentioned earlier is that there are multiple stages you need to take in your project whether it's generative AI project or it's something that would showcase a blend of machine learning traditional machine learning with generative AI you need to start with uh collecting information and data about it right doing proper research and then understanding what tools are going to be ideal for your use case then uh start with making sure that your data is in good shape. So any data cleansing, any data preparation. So there's a process that we call NLP, natural language processing. Uh we need to use or apply so we can prepare the data for machine learning training. Okay. So a thorough and wellthought um NLP or data prep-processing stage would showcase your strength in making sure that your system is rigid and healthy. And then document that's number four. Documentations is very very important. You have to be descriptive. Now obviously don't write like seven eight essays. No, be concise but descriptive enough where you are showcasing your thought process, explaining your decision making uh in case there there is uh there are two separate solutions that are ideal for your problem. Why did you choose solution A over solution B right and the trade-offs right between these two solutions? Diagrams are always useful because this showcases that you have the ability to understand complexity and be more like a data architect and system architect. Uh we'll wrap it up with standout signals. So having a unique and polished capstone project from end to end. uh making sure that you have any awareness of ethics or let's say uh barriers depending on your country or company's regulations or if there's a company that you aspire to join you can try to embed let's say you know Facebook's uh let's say user policy so trying to make sure that uh you have that in mind and it's being mentioned and well defined in your project where you do not violate any regul ations so you have clean tested reproducible process and even code because as I mentioned sometimes they may grab snippets of your code and try to test it and if it just doesn't run then you know they're going to lose trust on the quality of your work. Thanks for explaining all of that in detail and I think there is a relevant question related to this. So uh for some beginners uh they wanted to check with you how many uh portfolios uh would be required to as a proof of their work if they do not have any experience in the field. How many uh projects would you recommend them to work on? Yeah. So that question is going to depend on the size of the projects you're trying to accomplish. I mean if you if you have you can have one capstone project that can showcase multiple technologies and could be very comprehensive and complex and it's got a lot of components and that is just defined as one project right uh or you can have smaller projects and uh as long as those showcase all the points that we mentioned I think you're in good shape um the more the better always right so uh But having just one comprehensive project, try to take things one step at a time. That's my point. Don't overwhelm yourself bit by saying I need to build five projects. So you end up starting with three projects at the same time and then you don't finish those. No, finish one project uh at a time. And I think as you get more time and your experience grows, you will be able to start working on more advanced projects. Uh thank you and I hope that answers uh your questions as well. I think there were like two to three questions on that. Uh so before we move on I can see a few questions regarding the recording and the slide deck. We will be sharing that once the webinar is completed along with your certificates of participation. Uh so uh coming back uh to the topic uh so basel we've covered uh the elements that go into making a good generative AI portfolio but uh sometimes when you are transitioning especially uh it's important to have a certain set of skills uh to create a very strong portfolio and to stand out for your to your hiring managers. Uh so could you explain what are some of the key skills that you look forward uh to anyone anyone who's transitioning in this field and how can they work on these skills um and what are some of the ways in which they can build these skills and add them to the portfolio to improve their chances of you know getting hired or improving getting a salary hike possibly. So yeah. Yeah. So because we are in this AI boom, modern AI jobs, I would say expect a full stack understanding of multiple tools. This is because we're still trying to define what's good and what's not good. What is appropriate to a specific system, what's not appropriate to that system. and you have multiple companies that are basically building their own uh framework. So you have for example a really good example is when you uh start learning about deep learning you have two main frameworks pietorch versus uh tensorflow. So these two are completely different systems. Now they do achieve the same objective and they do help you build neural networks but the coding and uh the dos and don'ts the standards and a few things here and there or nuances that make them different and you need to be good at both basically just so you can have this full stack understanding per se right so from fundamentals to all the way to deployment and because now knowledge is accessible online and also there are a lot of courses we'll talk about some of the courses that are available uh at simple learn that will help you have this comprehensive knowledge and learning about these different systems you need to be wellversed in these technologies so um let's say for example we have a ragbased medical policy assistant that could you know generate advisory or advice or recommendation Right? So you need to know about the you need to have general uh what we call AI literacy. Right? So knowing when to use rag versus fine-tuning. So these are two separate approaches based on the complexity of the task. You know sometimes you don't want to use rag because it's not as complex as rag and rag could be costly and could also cause some uh security concerns unless it's really needed. So it depends on the scenario. We're talking about medical policy. So you obviously need to think about sensitive patients data. And then um the next stage is models and architecture. So you need to understand your current systems token limits any embeddings that are needed if you need to do batch normalization chunking all the steps that are part of preparing the infrastructure and the data and then next LLM applications. So um there are different workflows. The most common workflow you've got the lang chain workflow is highly recommended to start with and it's open source so it's accessible to everyone and then learning about different agents. So we have multiple agent frameworks. So learning about the agent orchestration, how they work, uh which one is more costly than the other, the pros and cons overall and then understanding the governance. So back to our medical policy example, you need to understand what we call PHI that is sensitive information and uh if there's any PHI filtering, if there is any masking, this is a process of hiding sensitive information but still making it uh um easy for the system to understand and provide results. And then doing bias checks. This is very important especially in healthcare nowadays when you want to make sure certain systems are not biased towards specific subgroup and uh in the population. And then finally with the deployment aspect, you want to learn about the difference between fast API and something like flask or you want to learn about how you want to deploy your system using docker where it can be compatible and um we call containerization. That is the process of just having a container for your system. Uh so it can work on multiple operating systems and environments right uh so um we also have a question uh for you again I think it's uh quite relevant to what we are talking about uh so there are some questions about capstone projects that we have highlighted towards the end and uh the question is if uh uh a capstone project which is completed using AI effective and also So there's one more question on the quality of projects. So when they try to work on multiple projects, it can impact the quality. Uh so how would that be viewed from a hiring manager's uh perspective? Yeah. So don't go if I understand the question correctly, we're talking about if what is the fine line between having too many projects and having like one or two three like a smaller list of projects. I say uh focus on depending on your expertise and where you're at in your learning journey. Start with smallcale projects just so you can build your confidence right now. Don't go too diverse with the project. Don't start on four or five different ones as I mentioned earlier. just start with one and then uh upon that one if it's if you're able to expand that existing project to uh more complexity and add more components to it more technologies that will be nice but if you now feel like you're ready to move on to another project different use case where you can showcase all the skills we talked about that'll be a really nice upgrade for your skills right Uh thanks for answering that question basis. So uh that also brings me to the next question. We've already spoken about the skills and we have a lot of participants who are working on their portfolios currently as well and they want some tips on how to improve that which we did cover. uh as you have also seen our learners create portfolios as a part of the uh programs that we offer at simply learn what are some of the common mistakes that you have seen our learners make or anyone who crafts a portfolio at first peaks and uh what are some of the tips that you can give them to avoid these mistakes if they are starting from scratch right now uh and they're also trying to transition from their field so we I can see some questions about people with a lot of experience over 10 years of experience trying to transition into uh generative AI and quite confused about how they can craft their very first portfolio and how to work on it. So could you share some uh thought your thoughts on how they can avoid a few common mistakes that you see a lot of participants or learners make? Yeah. So, uh this list that you see here could be really um I would say career limiting because uh some of these things if they are uh showing in your portfolio or in your work, your resume uh they could be they could be uh pretty much the tiebreaker between you and another candidate or could be like really not uh could take you out of the uh candidate. Y pool so to speak. So um first of all avoid having a tutorial style project. I think we mentioned that earlier in example is that don't do not try to copy paste a tutorial from uh LinkedIn or YouTube or uh something you found on Kaggle which has a lot of uh great examples but these are nice as uh let's say um nice for learning the basics but if you want to move from the weak portfolio label that we mentioned earlier to all the way to the strong portfolio that is a common pitfall that I see in a lot of applications. You need to address that, right? So creating for example very popular geni projects creating a Q&A bot using a YouTube tutorial. There's so many tutorials online there and if a recruiting I mean being in the in the position of a recruiter you see that project all the time from different candidates. What would you do? I mean obviously you're going to just you know roll your eyes and be like that is just typical right? So um the biggest killer also is lack of documentation and clarity. So not having a readme document. So this is like a a technical terminology. Providing a readme document is essential. now is not just for AI and generative AI applications but every program every project you work on in GitHub and you go online to GitHub and and take a look at some of the projects out there you'll see in 95% of the time someone especially for the great project you'll see someone that put together a readme uh document which is the highlevel explanation of what the program does how to operate how to run it if there there are any nuances or also system requirements. So it explains the problem it's trying to solve and it explains how you can run and operate your program. So if as a manager if I see this I mean I know now that this person is going to be ideal to hire for my team because they're going to provide proper documentation. They're going to be able to collaborate very well with my uh with other teammates. Right? So that's very important. No performance evaluation if you remember that was another factor we defined as uh if you don't have a performance evaluation then your portfolio is considered weak or lacking uh the depth. So you need to showcase that yes I understand my model cannot be perfect. Here's how I evaluated it. This is the drawback of the model. These are the mistakes or the errors it's making. Right? So you need to include metrics or let's say output uh for your evaluation. There's a process of splitting the data. We call train test split. So you split your data into training data set and then you have also the testing data set. And sometimes you have three uh segments. So you have train test and validation. So you need to showcase that process. And um I think the the the fourth point is very obvious unstructured and incomplete code. Uh make sure that you finish your work, right? Uh we're not talking about let's say you have seven eight steps that you're planning on working uh and you didn't get the time to finish the seventh step. Sometimes that's okay if you explain hey this is my next objective. But inside these steps, if you have code snippets that are just incomplete, they have errors and uh they cause noise, avoid doing that or avoid having that. And then the last one is not having real world relevance. Make sure that you have realworld relevance. And my recommendation here, there are a lot of industries right now utilize that utilize AI. So you have to I would say pick at least three four industries that you are really interested in and you think you have a chance of getting a job at and you know let's say you're interested in um technology infrastructure or interested in uh internet services or retail business like Amazon. So make sure that you have use cases that are relevant to these businesses or industries you're interested in so you can come up with a real world relevant example. Right. Um and I think you've almost covered all the mistakes that people usually make and uh uh in and we also have uh a question related to portfolios again. Uh so this um Kashish has actually asked uh how often should one update or revise their portfolio to keep it relevant. Yeah. So that also is going to depend on the type of project you worked on. So let's take it as an example. I mean your awareness of your work is going to help you gauge that pretty much because let's say you use a framework and you found out that eventually you know throughout your research being uh close to the industry subscribing to blogs that always share new technologies and you realize that there's a new version of that framework that you used that could solve for a specific problem you struggled with throughout the project or could really improve something that you thought was a little bit uh sluggish in your uh process. That's why it depends on you being so aware of the steps and the inner workings of your project because it'll help you gauge hey definitely I need to revisit and fix that uh problem or let's say update the system or let's talk about system incompatibility and especially in Python that is the most popular language we'll talk about Python shortly in Python we have because it's open source we have new versions of libraries being rolled out almost every week and some libraries are compatible with specific versions of other libraries. They have to talk to each other. So let's say when throughout your research reading you found out that uh the latest update let's say of tensorflow this is just an example latest update of tensorflow now does not work with an older version of numpy that's another python library and you know that you use that version in numpy so you have to go back to your project make sure that is uh still easy to run and you update the proper libraries and the specific additions or versions of your Python modules, right? Um, and I think we've answered that question. Uh, unfortunately, we do not have time to take up any more questions in between the webinar in the interest of time. Uh, so I'm going to quickly move on to the next slide on the career paths one can get into because tend to AI is quite a diverse field. What are some of the options one can opt um options out there for anyone who's transitioning into this field and um how are the salaries like for someone who's entering this field and is does not have a lot of experience. Yep. Uh so generative AI engineer that is becoming a very popular role and on high demand. Think of this as a person that would uh build pipelines that would help facilitate agents and uh rag systems and um you have also prompt engineers. These are less technical roles. So these are people that help with the testing. Let's say the gener system or a modified or fine-tuned LLM system. The prompt engineer is going to handle the testing and the quality check. So they will optimize the prompts, evaluate the models. So help them measure the success of these models and systems. And then you have AI project manager. So we're shifting more towards business now and less technical knowledge. So these are the people that do have uh the expertise in both realms. The business and its requirements, the business acumen and the some of the technical um knowledge. So they translate their business needs into AI features and AI capabilities. Then you have the uh applied AI scientist. So this is a person that is spending great amount of time in experimentation uh benchmarking pushing the envelope with new technologies and techniques and then on top of that helping with the optimization because at the end of the day you want to make sure that your system is easy to run on your company's production uh infrastructure. And then the finally is the AI solution developer uh or even automation engineers. These are the people that help with the enterprise automation. Uh it could also be u some blend of what we call dev ops. These are for development operations engineers. They help with the proper libraries that are available for your system and the production environment. And when we say production for the ones that are not familiar, this is the environment that the business operates and presents to their end users. So this is very important. uh tier in your ecosystem and you know you want to make sure it's successful. So AI solutions developers or automation engineers are responsible for making sure that the whole uh or the full cycle of the AI system is working well and operating efficiently in production. Uh thank you for sharing all of those insights. I can see a lot of uh people in the chat asking about the attendance form uh and also the certificates recording. uh so I just wanted to let all of you know that we will be sharing a form towards the end of the webinar and uh we'll be covering a few more slides so you will have to wait once till that gets completed and you will be receiving your certificates recording and the slide deck to your email id and I hope that clarifies that so it would be great if you could all like concentrate on the webinar here and refrain from chatting as much right uh so now uh we've also got a lot of uh questions from the audience about uh they are new and they do not really know how to get uh or make a great portfolio using AI and they do not really have a structured learning path and I think that is also where uh the advanced executor program in applied generative AI comes in and uh this is offered by simply learn and IT Madras Praerta course a lot of you might already be aware of it uh but as Basil is also a generative air trainer with us I believe it would be great if uh Basel could uh if you could actually um explain the core topics that are covered as a part of the curriculum because we have already covered these in the previous slides as well. Uh so um it would be great if you could just let the audience know the kind of topics that we will be covering as a part of the course that we are offering. Yeah. Uh so I think this is a a nice way to tie all the points that we uh mentioned earlier together and make sure that we have an idea about how u an organized course could really help you throughout this journey. Especially if you are pivoting in your career or you're starting from scratch, you don't know where to start. This is where an organized course or cohort could help you and guide you through these steps. So first of all we I mentioned you heard me mention Python here and there and this is a very very important module although we say that this is optional just because a lot of you may have some background in Python and you might find this a bit repetitive but for someone that's starting from scratch highly recommend taking that and even for the ones that have some knowledge in Python honestly you know you want to take the this first module because you always learn something new you always learn about different techniques techqu. So think about writing clean, modular, testable Python code. It's very important. Again, back to showcasing your project elements and how uh clean your project is. So it turns your Python skills into think of this as a power tool in terms of automation, working with different APIs, working with different frameworks. So it's not just building a a standard script but just real AI workflow engineering. And um the next module is AI literacy. So you definitely need to have an introduction to the world of AI. So uh this is more of a like conceptual foundations uh type of course. So you'll learn about how LLMs actually work, what engines they use. You'll learn about uh what we mean by tokens, embeddings, the NLP style that I mentioned earlier about pre-processing the difference between um different AI types. I always see people you know there's a lot of misconception there uh between the different definitions. So you have machine learning or traditional machine learning, you have deep learning, you have uh generative AI and then when to use fine-tuning versus rag systems uh and and the drawbacks of using specific uh elements. So this is very important and uh next you'll be ready after finishing AI literacy you'll be ready to uh dive into advanced generative AI and the different models and their architectures. So you'll start with what we call the transformer model and transformer is an advanced um sequential so to speak advanced uh neural network and basically it's the engine behind something like chat GPT and then we'll once you're comfortable with these different architectures of neural networks especially for transformer we'll look into large language models and uh we'll look into scaling uh model families like you have GPT, Llama, you have Gemini, Claude, all these technologies you hear about and use on a daily basis and some fine-tuning techniques. So things like a technique that is very popular, PEFT or Laura, these are two popular techniques for fine-tuning an LLM. Uh embedding the models also vectorization. So you can use that for data management. So that that's what you'll learn in module number three. Module number four, you'll start you'll continue this journey with advanced generative AI with building LLM applications with real world scenarios. So um you want to learn about prompt engineering patterns so you can tailor that engine around what you expect the users to put as a prompt and uh building what we call multi-step LLM workflow. So learning about the architecture the nuances of the different architectures um ingestion chunking embedding. So all of these techniques and what we call rag components also something that's very important in LLM is learning about query transformation. So you have what we call the process of reranking um context compression right I'm throwing a lot of terminology at you but hopefully this gives you the scope of how um comprehensive and diverse that the course is going to be. So all of these components will help you understand how you can build real applications using these different system lang chain and different orchestration systems and why you need to have a diverse tech stack. Okay, so I know we're uh getting short on time so I'll recap module five and six quickly. uh these are going to be learning it will help you learn about additional capabilities in um genai. So think about image generation capabilities. So it's not just for text generation image generation. So you learn about different models like stable diffusion, their fundamentals, uh GAN architectures which help with generating uh images from a prompt or text and then uh with module six you'll learn about MCPS or agentic frameworks. So this is becoming a very popular topic. MCP stands for model context protocol. So these are going to help you understand how you can leverage agents for specific task. Uh how you can do what we call the tool calling the planning of the uh MCP doing uh multi-step reasoning and monitoring that whole ecosystem. And then we'll wrap it up with module seven and 8. Module seven is important because this is where you can showcase your um your organizational awareness meaning not just organizing your time but the organization itself. Generative AI governance is important because you want to learn about safety bias uh components. You want to learn about the concept of hallucination in LMS right or providing misinformation. How do you handle these systems and how do you make sure you are documenting that work so you can showcase that work in case there is any uh auditing or investigation right and then finally the capstone project think about this I talked about all these separate components from module one all the way to seven you want to have a project a comprehensive project that could wrap up all of these concepts together into one. So expect a very intensive full end to end AI product for that capstone project. Thank you for uh explaining the curriculum in depth Basil. I think everyone has a complete understanding of that right now. Uh also uh because we are in short of time we'll quickly go through the hands-on learning experience uh that we offer at uh simply learn as a part of the course. Uh so basically you get 60 plus hands-on exercises which means that you're not just learning tools and uh you also get seven industry relevant projects and also you'll be working on end toend applications. Uh could you point out any other highlights uh as a part of the uh you know participants can benefit as a part of this course. Yeah, I think uh building seven industry projects that's a big one for me. Uh especially at Simply Learn, we try to use something that is unique and uh put together by multiple experts. It's not going to be just a standard typical project that you can find on Kaggle, for example, or YouTube. It is going to be a very well-put and comprehensive project that would help you understand things like the lang chain pipeline, rag systems, uh building AI systems and just real world use cases and these can help you with your AI portfolio online and it if it doesn't encompass your industry, it'll at least give you an idea about how you can apply what you learn from it into a different indust. industry right uh thank you for saying that so um I will quickly take you all through the tools covered and the skills covered because we are almost um uh the time is almost up I'm not going to elaborate on those uh so we will be covering 19 plus tools as a part of this program and you will also be uh getting 13 plus skills will be covered as a part of this course we have already spoken in depth about these skills in the previous slide so you already know um how you can add these to your portfolio and how it can benefit you as well. So we'll be covering prompt engineering agent AI frameworks genai application LLM architecture and a lot more which Basel already spoke about. So I'm not going to get in depth uh into all of these things but these are some of the advantages that you get as a part of this uh program. Apart from that uh here are some of the industry projects that you will be working on. Uh so you have a a good idea of the kind of experience that you will be having and a lot of you had questions about what kind of projects can you build with AI. Uh so you will be enabling AI powered business intelligence for organizations. That's one of the projects. Another highlight is building a Python adventure game with GitHub copilot and you can just imagine the impact it would have on your resume. And some of you had questions on the number of projects that you need to have. So you will not be working on this by yourself. But you will also have our mentor supporting you and giving you feedback and assistance while you work on these projects. Apart from that you have crafting an AI powered HR assistant and analyzing customer orders using Python. And also we'll be covering charg based storytelling. These are just some of the examples. There are many more projects capstone projects as well but we will not be able to get into all of that in detail. uh we will share the uh link with you in the chat course link with you. So you can go and take a look at it by yourself right and apart from that uh we also want to highlight the career support that we are offering as a part of this course. Uh so you will be receiving group mentoring and networking opportunities. So you will be uh you will be added to a slack community where you will get to interact with your fellow peers and the mentors. And that's also how you will be receiving feedback on these projects that you're working on. And one other aspect that everyone will be worried about once they take up a course is how you're going to appear for the interview. Uh so our team will be helping you with interview preparation and also assessments to improve your performance uh and help you get hired uh faster. Right. Um and apart from that we also have AI powered profile optimization. So again that's something our team will be working with you alongside as you complete this course. We'll be optimizing your resume helping you optimize your resume and your LinkedIn profile help you add the skills and also uh give you feedbacks on what might be lacking so you can improve your uh resume as well. And this is already covered mock interviews and mentoring. This is one of the highlights that we offer at Simply Learn uh is that we have great mentors like Basel who will be supporting and assisting you throughout uh to ensure that you are all set for your very first interview or for that salary hike that you have been looking forward to having. And that brings me to the eligibility criteria. I won't take much time. We'll wrap this up in like two or three more minutes and we will get to the certification part for which a lot of you are looking forward to. So eligibility criteria for this course is you need to have a bachelor's degree and an average of 50% or higher marks and it's great to have a basic understanding of programming concepts and mathematics and preferably it is uh good to have two plus years of professional experience even if you don't have it it's not like a mandatory criteria you can still uh join the course and uh a few of you asked about the fees of fee structure for this course so for Indian learners it is 1 lak 40,000 and this is also So because it is from IIT Madras Praartak you know that it's a reputed organization and institution and it has a credibility on its own right so you also have the option to pay in installments as low as 6269 per month and for global learners the course fee is $2,199 and you also have the option to pay this in installments and this is the certification that you will be receiving from IIT M IIIT praert so certificate of completion from simply learn and IITM and apart from that you will be receiving a Microsoft as your certification. So these are the basic um things that we wanted to cover regarding the course and I will also be quickly launching a poll for anyone who is interested uh to join the course you can go ahead and click on a yes and our team will reach out to you. Um if you have questions also you can just click on a yes for that because um unfortunately again with the interest of time we won't be able to answer your questions here. It's a good opportunity probably by the end of uh by mid next year you will be able to transition into generative AI with all the projects and portfolios that we covered during this webinar. Right. And I would also like uh to take up a few questions in the meantime. We will close this poll in another in a minute or so and we will launch the participation poll as well. Uh so you guys uh can do not have to worry about that. Um I think we have a few questions regarding a lot of questions regarding generative AI actually. Um okay which platform uh which uh Basel this is a question that we've got from one of our learners. So which platform is best for creating a generative AI portfolio and what would you suggest? Yeah, it depends on what you're trying to achieve here. But I think uh my recommendation is start with lang chain. Um they have a lot of great tutorials online and you'll find uh also that in our course we focus on that because it's one of the most common frameworks nowadays. So I recommend uh using or starting with that. Right. Okay. All right. So, um I think uh we have received quite a few responses. Um I'm going to keep this live for another 30 seconds. If anyone's interested, you guys can go ahead and let us know in the you can respond to the poll. Uh a few of the questions regarding portfolio has already been answered. So, I'm not going to go into that right now. Okay. Uh just give me a minute. All right. So I can see a lot of responses. Uh there's one more question I think for you Bassel. So uh there's a person who wants to know which field in generative AI uh has the most potential to grow by 2030 and what are your thoughts on that? I wish I could tell the future. Um but I think the fact I would say this though the fact that you are starting this journey today uh you're basically doing step number one which is the most important step. Um and what we mentioned earlier staying diverse in terms of technologies. Don't limit yourself to one technology because let's say you know you learn about three four different applications or frameworks and one of them ends up being the winner and the one that's going to be uh the standard for the next 10 years. Um I say that would be very useful for you that then you can focus on that specific one. That's why we always ask our students to learn about have that full stack mentality meaning learning about all these different technologies and frameworks. Um again it's hard to predict the future but I think what it would be in the future mainly something that would deal with a lot of optimization and making sure the system runs smoothly um and uh a lot of rapid prototyping. If you are good with that skill, I think it's not just relevant to the technology, but it's also relevant to anything really. So, it'll help you be relevant in every decade pretty much is if you're able to uh follow that process we talked about from defining the objective all the way to doing thorough planning and architecture building all the way to the evaluation aspect. Right. Uh thanks for answering that uh Basel and I think uh we've got a lot of responses. I have shared the link course link in the chat. So if you guys are still interested you can go ahead and click on that and you will still see an option to uh reach out to the team there. So I'm going to end this poll here and I'm going to launch the particip uh the poll for certificate of participation right now. Uh so you guys can go ahead and fill up your full names. Uh I saw a lot of questions regarding what you will be receiving. So you will get the slide deck, the recording for today's session um and the certificate in your email. The timeline will be 24 hours to 48 hours. In case you miss out on filling up this form, you might not get your certificate. In that case, you will have to mail us at [email protected]. And also while you are filling up your form, I would like to take this time to uh actually thank Basel for you know joining us today for being such a wonderful uh speaker and sharing such amazing thoughts and insights with our learners. It was a wonderful session and I thoroughly enjoyed being a part of it. I've got a basic idea of what I can also possibly do uh to make a great generative AI portfolio. So thank you for being a wonderful uh speaker and um I also have to ask you one more question before uh we close the session. So what's one piece of advice that you have for all the participants of today's webinar? Yeah uh first of all thank you for having me and definitely enjoyed the conversation and the questions that were asked on the platform. So, I I definitely appreciate that. And my message to everyone honestly is that you guys are doing the right thing by asking questions and uh you you will eventually get to the bottom of it. I promise you. I know things may seem overwhelming, but the fact that you are here and asking the question, what's next? What can I do? At some point, you will achieve your goal. Uh so, just stay patient and um perseverance is very important. uh take things one step at a time. Don't overwhelm yourself with so many things and uh I think you'll be set. Thanks for that great advice, Basel. And I also want to uh tell you all that you've been a great audience. Thank you for all the engagement and participation from your side. Uh and um I'm going to quickly end this poll for the certificates. So like I said, in case you guys have missed out on this, mail us at webinars.net. Thank you for being there. Thank you for participating and if you want us to host any more webinars, you can always reach out to us and let us know your thoughts and we will organize uh really good webinars for all of you and I really hope to see you all in the upcoming webinars. Thank you everybody. Hope you all have a wonderful day ahead or a good night ahead. Thank you.