What Is Vibe Coding? | How Vibe Coding Can Help In Your Career As A Software Developer | Simplilearn

to have you here. Kaustubh. Thank you for joining us. Okay, every Hey, Anshu. Hi, Muhammad. Hi, James. Thank you for joining us from California. Hi, Dileep. Joining us from Rajasthan. You're a freelancer. Hi, Anthony. You're joining us from China. Wow, great. Hi, Eddie. You're from Dallas. Hi. Someone from Makalia. From May She's here. Great. Amazing. So, we are also live on LinkedIn and YouTube. So, A very warm welcome to all of you guys. One second. Let me just check my screen share again. Apologies for the interruption. Okay, so yes, my LinkedIn redirected because I was streaming this live. Great. Okay, so now that we have everyone here, we've done our round of introductions. I'm very very excited to begin the session. My name is Ananya and I would be hosting the session on behalf of Simplilearn. We have an amazing 1 hour plan, so please do stay tuned with us. We have an amazing guest as well on the panel. But before we begin, a couple of ground rules for the session that we need to follow. 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, so we might end up missing yours. So, please use the Q&A box effectively. We will try to answer all your queries during or at the end of the session. We have a dedicated Q&A segment planned. Secondly, avoid sharing any external links or personal details on the chat box and keep conversations only relevant to the topic. We will be monitoring this strictly and any irrelevant messages will be deleted immediately. And 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 webinar, we will provide an attendance certificate that you can share on social media, add to your resume and even update on LinkedIn. So, please do stay tuned till the end of the session and participate in the poll. Unfortunately, we won't be able to take any manual entries. So, please do engage in the poll towards the end of the session. Right. Okay, so we have a huge bunch of audience here and I'm curious to know has anyone here attended Simplilearn webinar before or is it your first time? Please do let me know in the chat if it's your first time. You can just type first time and let us know. We're very you know, eager to know how many of you are here for the first time. Okay. I guess everyone needs a little bit of time to chat to type in the chat. Okay, a couple of you I see that you've posted in the Q&A box. Um That it's not your first time. Some of you are saying that it's your first time. Okay, a couple of you are here have attended multiple sessions before. Thank you for being here again. Thank you. Okay, so allow me to quickly brief about Simplilearn and what we do. We provide certifications and career aligned learning paths in various categories like data science, project management, Agile and Scrum, business analytics, generative AI, software development and a lot more. We have helped 8 million plus people in their career growth across 150 plus countries with more than 50 partnerships. And like I said, you know, we have a great ecosystem of great partnerships. We've built a curriculum in collaboration with highly ranked institutions around the world and also global tech giants like Google, AWS, Microsoft and a lot more. So, these partnerships are really shape what we teach and how we teach it, which means what when you learn with us, you are learning exactly what the industry expects and requires. Our learners report an average 50% salary hike. We maintain an 80% graduate rate, which is much higher than average industry average for online learning. We also have a rating of 4.5 from our community of learners. This All of this is primarily because our programs are live, hands-on and taught by practitioners and not just academicians. So, this show kind of learning experience and dedicated guidance like that we've been able to provide to millions. With that, we can you know, get on to the you know, main discussion of the But you know, before we get to the main topic and I hand over to our amazing speaker here, I would like to know like most of you would agree that a software developer's role has changed a lot in the last one or two years since AI usage has increased so much, right? So, can I can I get like you know, thumbs up or yes in the chat if you agree that software developer's role has changed a lot? Yeah. Yes. Yes. Yes, I see some reactions coming in. Great. Great to know that we're all on the same board about that. So, the role of the developer is definitely rapidly evolving and many industry are probably just beginning to realize this transformation, but it's not all bad at all. If you see some data points here that we've compiled here, four out of 10 developers acknowledged that AI already opened new career opportunities for them. What's even more striking is that in 2026, seven out of 10 developers can expect their roles to evolve further because of AI. These changes, these statistics and trends highlight a significant movement in software development domain as AI continues to redefine how developers work and what skills they need. So, let's explore what this means for the future of development and how we can adapt to these changes. Okay, so in the next 1 hour, we are going to look at why coding and how AI has changed development practices. We'll discuss what tasks it's replacing versus those it's augmenting and showing where we understand where our human expertise still remain irreplaceable. Next, we will also highlight the skill profile of a high-value developer in AI-driven environment. Security is another vital area we will address. Despite AI's advancement, there's a security blind spot that developers must still own. And finally, we'll gaze into the future discussing next 12 to 24 months world for developers and wrap up with brief overview of one of our programs. So, by the end of the session, you will have a clear understanding of AI's role in software development and what it means for you. Before we begin, thank you, Kaustubh. Thank you so much for your lovely words. Great to know that you know, you have attended more than four webinars and you have learned a lot. Thank you so much for being here. Great. Okay, so we have a lot to cover today and to guide us through all this, we have a brilliant, dynamic expert, Arpit Gupta. Arpit, lead engineer at Google, brings many years of invaluable experience in building large-scale mobile and software systems that millions of users rely Currently spearheads initiatives at Google that focus on developing, scaling applications and tackling complex data and performance challenges. Before his time at Google, Arpit worked as a senior product engineer at GoJek Tech and as software engineer at PhonePe where he contributed to high-growth consumer platforms and built production grade applications at scale. With Arpit's depth of knowledge and experience, I think we're in excellent hands to delve into the session, but before that, Arpit, a very warm welcome to you. It's great to have you here and to host you today. It would be fantastic if you can add on to that introduction and say your hi to all our participants. Thank you, Ananya, first of all and thanks to Simplilearn. Hi, everyone. I think Ananya, you covered about me. I think beyond I would I can cover. So, I think thanks for you know, having me over here. Glad to be part of Simplilearn and you know, as we go ahead, we'll I can you know, provide my insights around how the AI and software development is evolving in general. So, yeah, looking forward to Perfect. Perfect. Great. Okay. So, before we begin, I just wanted to let everyone know that as all of you taken some time, you know, today to be with us today, we want to ensure that you get the most of this session. So, in addition to the amazing insights that Arpit, our expert, is going to share, we've also put together a resource for all of you who are here in our webinar. So, you will be getting a live coding startup playbook which includes like, you know, prompts that write production-ready code. You will be getting this in a PDF format on your email after the webinar. So, please do stay tuned till the end and engage with us more to ensure that you get this resource for free. Okay. Great. We can I think there might be some disconnections. There. Our host. Hey, we are back again. You are on mute, Aranya. My apologies. I seem to have had a network issue. I will We can write where we left off. Okay, I hope my screen is visible. Yes. Okay. Okay. So, I'm not sure till what I was audible. So, I'll just like, you know, sort of repeat my question. So, Arpit I was talking about how when you started your career as a software developer, AI must have been like, you know, probably a term that was used that we discussed about that we know of sort of but not really in detail. But today, like, you know, every day we use AI tools and platforms increasingly across like, you know, different fields and different domains. We saw like, you know, how widespread the usage of AI has become amongst developers or in the process of development. So, can you start this session with, you know, sort of painting a picture of before and after? How was the traditional development era like compared to what it is right now? Surely. Surely. I think, you know, anyone who starts their software engineering journey in the past starts with learning the basics, syntax, you know, couple of understanding around frameworks, understanding around debugging and all that stuff, right? You know, all this AI started coming up in 2022, 2023, late 2023, right? Where we started all this with the chatbots. But eventually, now we realize that it's not just about learning the system learning these syntaxes, frameworks, but rather it's about your role is now more about builder rather than understanding all these syntaxes and all, right? So, I think in general, it's it's expected that a software engineer becomes an architect in a way where you have a cross understanding around different systems, how these systems interact with each other, rather than being an expert on one of these systems and all the details. I think for sure like, apart from the development becoming faster, the execution becoming faster, I think it would be not a bad idea to say that, you know, if you are writing still writing code by hands or still writing syntaxes, you are kind of, you know, somewhere behind in general. So, I think that's how all these are evolving. These are very, you know, quickly evolving. And we have as a software engineer, that's our energy is that we should evolve as fast as well. Right. I think that sets the perfect context for us to discuss about live coding because, you know, with the advent of AI, there are newer and newer terminologies and frameworks and concepts emerging every other day. And the latest fad we have now is live coding. I must say just a few weeks ago, we did a live workshop on live coding and my word, the response was fantastic. There were hundreds of people in the workshop wanting to learn how to live code and were genuinely enjoying it, too. So, if If anyone is into the benefit of our audience is here, so can you take us through what live coding is all about and what its impact is on software developers' role? Sure. I think um Pre-the live coding or AI-assisted coding, you know, as a as a co-partner with you where you have your ideas in mind, you have the instructions jotted down in your mind how will you be doing it. And there's someone who will assist you do it. Who knows entire um best practices around those coding styles. Who knows uh best about syntaxes. And uh someone who can understand multiple, you know, files, modules together at the same time, right? So, that's where I think um this live coding has evolved to a assisted uh it is assisted coding, you know, partner you can think of. Where, you know, rather than like coding from scratch, you actually just have to articulate your requirements in a right way. The more right you articulate your requirements, the more best results will be getting out of that. So, I think it's about like processing your instruction to generate, you know, and refine your entire code base. Okay. Perfect. Great. Um so, let us like, you know, digress a little bit, Arpit. Working alongside AI has become a norm for all of us now. So, specifically for developers, what are the skills you think they need to work on to keep themselves up to date and evolved, like you said, and relevant in 2026? Correct. I think one of the thing obviously we all know is like we all know this term which is like prompting engineering. I think it's an I would say it's an art to prompt even. The you know, the the quality of your prompt decides the quality of the live coding outputs, right? And there are a couple of things which are being evolving when with with these live coding and AI agents coming in, right? Now, there are autonomous agents which can not just So, it's about you can So, earlier like few days back or few months back, it was about, "Hey, you give some prompts, the the agents will do some work." Now, these autonomous agents can eventually do the larger amount of work. They retain the context of the work which essentially means a medium a medium to, you know, medium high efforts of task can be done asynchronously given you, you know, given you like literally dissect the problem statement in a in a in articulate in the best way. So, that's how these things are evolving, one. I think the one of the thing which I've also observed in my experience is live, you know, prototyping or uh prototyping has become very quick, right? You don't really have to wait for your PMs, your your UX teams, your your counter engineers to, you know, to align. You can quickly go, prototype, show, and evolve. I think that's where the the process of ideation to landing is becoming very, very short. So, I think with live coding, not only you you have can improve your productivity and output, but also given you can get back that time to learn something new and more. I think that's how I would really put it. Great. Thank you for taking us through that. I hope, you know, everyone noted that because like that would be very crucial for anyone looking to enter or grow in the software development field. Now, in this next segment, we're going to talk about tools and, you know, the workflow a bit more in detail. And I'd like to ask the audiences here like, you know, whether you're a coder or a non-coder, do let us know like, you know, what tools, AI tools you like, you prefer, and use the most. You can type them in the chat and, just let us know your responses. Cursor, Copilot. Yes. Anyone else? Thank you, Naresh, for your answer. Anyone else? Cloud, Codex, yes. Uh Claude was something that I would have told as well if I were you know participating here. I had to do a lot of analytics work today for a presentation and Claude really helped a lot in seeing patterns and doing some calculations. But yeah, from more of a technical standpoint, I think I see a lot of people saying Cursor, Claude, Copilot, Codex, Open Code. Notebook LLM, yes. ChatGPT, Gemini, Google Studio is great. Okay, a lot of a variety of tools that have come in. Um what about you, Arpit? Like what are your What are some of the tools that are your go-to? Obviously, being a Googler, I use most of the tools and internal tools of Google. And Gemini is my go-to for anything. The one we we use is like like does almost 95% of my job, daily job. It automates 95% of my daily job, be it from coding, from from design understanding, from designing the systems, from architecture, and all that stuff. I personally couple of time have used couple of these, you know, mentioned uh tools as well, like Claude, ChatGPT, Cursor. But mainly I I go, you know, Gemini is my go-to for for all the things. In terms of like coding and all, they are obviously couple of IDEs which involve all these external plugins with these ChatGPT, Cursor, Claude. I think those are also very helpful. If you are in terms of app development, Android ecosystems or the iOS ecosystems, already have the inbuilt AI toolings to assist you with your coding. Yeah. So I think these are the set of tools obviously, you know, you mentioned like couple of folks mentioned here. Uh Cursor, Gemini, and I think these are the set of overall tools. I think more than tools, Notebook LLM is one of one of my favorite to go and you know, give all the documents I have and try to understand, "Hey, give me these detail out of all the documents I have given." And it will search out and then, you know, give me the best out of So I think I I've been using Notebook LLM for my personal uh you know, stuff as well. Have been very very great over there. And like what are some tools that and platforms you would say are very necessary for developers to know about and like you know, to sort of like you know, it's it's a must-have in their resume. What tools would you say I think uh There's something called as skills. A skill is something which is beyond prompting, let's say. If you have uh you know, a a repetitive task to do, you should not Let's say if you are prompting for one type of task and if you have another similar type of task, you would prompt another time, right? So rather than prompting another time or like if there's a need to do some repetitive kind of work, you can create a skill out of it. A skill is a instruction specified instruction set which any of these LLMs model can understand, can plugin, and they will only and only then act based on your skills which you have added in them. I think this is something which which you know, if if people are not using, they should really go and explore skills. What skills are, what skills.md files are, how to create those skills. There are tools in the market to even create those skills. You don't have to really manually type those skills. You just give prompt to create the skill out of that prompt and use the skill every time so that you don't have to write prompt every time. And you can specialize your skill in a certain way that it can it does really autonomous to the Then it's it's it's Then your LLMs uh then then interaction with the LLM does not you know, limit get limited to prompting and also or the best prompting also. Then it becomes literally like a chat. Mhm. In the background, it uses your skills to understand what it needs to do and does it. Great. Okay, I think like everyone learned about a new tool today. That's exciting. And yeah, so I'll move on to the next segment here. So Arpit, like what I'm going to request you to do is sort of you know, break down the development workflow into five steps and you know, maybe like walk us through the role of AI in each step. Got it. I think uh This more or less has not changed in the AI world but has evolved. The development stages earlier as well were around, you know, framing the problem statement, generating the you know, generating the structures out of it, then coding, then debugging, defining, and then testing and deploying. These development cycle or stages still remain more or less same. Which is like Let's talk each one of it bit by bit, right? The problem statement earlier we used to define or the frame, right? Which we are calling here is exactly the same but instead of having documentations, you can maybe create a document which can be served to your LLMs which can help in the in your next steps like scaffolding. Scaffolding is creating the project structure out of it, right? Your LLMs models skills can help you create those basic structure boiler Not even just boiler plate code but the basic structure of the entire code base or or the you know, project plan. Then over and above of that, which is usually our third step, so it is creating the code, right? Generating and all that stuff. Generating the code now and like as we all know like it's about prompting, skilling your LLMs and we express the you know, idea of project which is like the first step, provide instruction in the second steps in the first steps combined to generate the code in third step via LLMs, right? One of the interesting use cases I have been doing in general in my personal and projects is apart from coding, I have been using these frameworks to increase the coverages of tests. Right? We don't really have to bang your head writing testable modules, debugging them, refining those tests to the best. I think LLMs have been very very helpful around those areas. LLMs also understands the gaps in your systems, the gaps in your architectural pattern. And I think there is obviously you know, a feedback loop if something is not working very fine, you can any day go back and you know, come and back, "Hey, this is you know, the instructions which was supposed which was given and you didn't really work around that." So any any point in time, you can go and provide the feedback loop. So this is something which is like very useful to define your prompts, define the understanding of LLMs, define the depth uh you know, of the implementation details or the thinking process of LLMs, right? Before they even actually implement. The last part which I touched upon which is like your your unit tests or CCD pipelines or configurations, all of these are actually very very easy nowadays to learn through LLMs. One of the other thing in general which I really use these LLMs for is to understand the unknown code base, right? The learning curve has really really the time learning time has actually reduced from few days to few hours, I would say. Uh you can just onboard to any of the new code bases and you know, you can just go and explore. Try to understand, "Hey, what does it select the code base? What does this code base really means?" So yeah, I think the the area of expertise or the lines between area of expertise is also thinning out, right? So I would say like these are the overall workflows which I use in my day-to-day life, you know, with LLMs and and these recent big models. Okay, thank you for that overview. That was pretty comprehensive. And moving on, we have You touched upon this a little bit. Prompt engineering has become you know, it's it's slowly becoming an essential skill for all tech professionals and even for marketers and analysts for that matter. So can you share some prompting tips with us? Like how is a great prompt structured? Perfect. I think let's start with one of the you know, examples shown here and then we can we can deep dive more into details. So as you see in the left, which is like the maybe the idea is just about creating a login page. if you don't give the details about that, what do you exactly need, which coding language do you need, these LLMs or agents can create those in their best of knowledge, which may not be really useful for you. Right? You might be wanting to create a coding page in Android, but who knows, you know, if you're not giving the details around those, it might have created those coding pages in React or in iOS or in some other languages, The details in your prompt or in your skills uh which we have talked about is a must. So, you should add all these details. One of the things that I've seen which works best in prompting is a structured prompting. Which essentially means uh you can break your prompt in a step-by-step manner. Hey, you can literally say, "Hey, step one, do this." Unless the step one finishes, don't move to step two, step three, step n. That's one way of doing it. Second way is like keep the data very structured. Let's say you want to say, "Hey, this is my uh you know, architectural pattern of the code base. Follow this architectural pattern to create your new code base, right? So, I think you have to be very, very explicit in nature while you are designing uh you know, the prompt skills out of And uh you know, giving it to LLMs to generate something. So, I think the more the precision in your statements the more unambiguous your statements are I think that will really, really help you to get the best out of these agents. Great. Thank you so much for that uh you know, example and that illustration. Uh before we go further, I quickly want to um take a you know, take a minute to pause and flag something for everyone here because we have an amazing demo webinar that's coming up um that I think will really complement what we're discussing today. Uh so, the topic is how to build an adventure game with Python and GitHub Copilot. Uh we are going to guide you through a demo of the development of a text-based adventure game in Python uh with GitHub Copilot as an active coding partner throughout. Um it's going to be an engaging session as well as like, you know, technically substantive. Um it's going to be led by one of our instructors um Amy uh Saldi Saldia Ga who's also a self-taught AI ML engineer and Python and GenAI specialist. Um it's happening on a Saturday because a lot of you had requested us to do weekend sessions. Um and it's going to be all at 8:00 p.m. IST which is 10:30 a.m. ET. Um so, you have the QR code right here to scan and register. And I'm also just going to share the direct registration link um on Zoom uh chat here. So, this would make the process um easy overall. So, yes, we hope to see a lot of you or most of you um in the next webinar as well. Okay. So, I'll move on to the next segment. Sorry, before we move on, I think you said that you will be flashing out this QR code on the screen, right? Yeah. The one I'm seeing is I'm not sure if everyone is able to see it. The one I'm seeing is like a grayed-out patch of the screen. Okay. Okay. Thank you so much for flagging that. There must be some issue at my end. Just give me a second. Apologies. Thank you so much for flagging that Arpit. We have the QR code here and I've also shared uh the link to register um with everyone on the chat. Okay. Great. Okay. Uh moving on to the next segment, um I want to ask like our audience is here like is there is there something um um you know, particular um uh action or an activity you feel like AI cannot achieve and still requires a lot of your involvement and a lot of your uh direction. Is there anything that you uh you know, felt like that? Do let us know in the chat. Um I think it'll be um interesting to hear your responses. You can also use the Q&A box to let us know um what you think um is still requires a lot of effort from your side uh to make AI platforms or AI tools work. Is there anything like that? Okay. Maybe maybe not. Uh so, I think like you know, uh it's good that we have this uh topic here and we're going to discuss uh that um in the next segment. Um Arpit, so if I have to put it this way, um AI is doing two things. Um it has certainly helped replace um certain routine rudimentary tasks. Um and on another side, it has also aided in enhanced how we perform certain tasks, right? So, can you um draw the difference for us in the context of software development, of course? Um this is so that our audiences know exactly what areas they need to focus on and hone their skills further. Got it. And I think this is you know, the topic which I'll be touching is also uh you know, one of our question in the Q&A as well that uh you know, is it still important for a learner to understand syntax, learn the syntax, understand the basic concepts like memory management or complex algorithms and all that? It's from uh Mukarala Sai Satya. Uh I think this is 100% must for engineer. You should know basics about the software engineering. What does memory management means? What does even in even in the AI-assisted coding environment, you should know all the basics of concepts of computer science, right? Given AI does most of the job I would say pretty well, 50-60% of the job. But it also hallucinates. It also gives you wrong output sometimes, right? And that's where your learning around these basics, fundamentals will help you debug it. Right? So, how I think like which AI is not really doing good is around AI cannot at least for today AI cannot do very good in the very large system designs and architecture. Right? They are helping there. But it still has to evolve. Right? But obviously, the basics of code generation, the basics of integration testings, the basics of unit testing, the basics of scuba testings. The documentations is also somewhere is helping you out. The creation of documentation out of your code is something like it does really well, Syntax and all those error detection, I think are were already been there before this AI world itself. Uh because like with with the these new advanced IDEs in place. But obviously, these got enhanced a much better way. They the current IDEs assisted with AI agents gives you not just the syntax error, but also tells you about what are the best practices you can follow. I think there these are some set of things which still we should you know, we should we are getting help with AI. But the one thing which we as a software engineer should learn, evolve is system designs. Is is security reviews, right? Is how the the parallel programming or you know, the parallelism which AI has done. Can't really work. Basics of business knowledge. Right? Ultimately, what we are trying to do is like asking AI to put some business logic, right? So, that's where it is a partner for you. But your brain is the one who will be doing most of the your thing, right? Treat AI as a collaborator rather than you know, someone uh I've heard this from a lot of people in general that, "Hey, is AI replacing uh you know, software engineers?" Not really. Your job will evolve. We you as a person will evolve. Uh and what I feel is like it's about assisting together rather than you know, being competing together. So, yeah, I think these are the set of things which AI is helping us making or bridging the gap between or bridging the gap of time between thinking and execution and you know, launching. So, I think AI can do couple of work. But whenever there's a creativity, where there's an understanding of systems and all human intelligence still needed and will be needed always. Right. So, you mentioned the hallucination uh Arpit and I think that sets like the right context to discuss about this. And my question is also around that. Um so, you know, when we talk about AI, we cannot talk we cannot not mention not discuss about AI's hallucination. um how much it requires security verifications like mentioned in the previous slide. Um so, if someone here is using AI tools to code, what are some aspects they need to be extra careful about like what are some potential mistakes and weaknesses that AI generated codes are likely to have. I think first and foremost thing is the generated code should be reviewed and reviewed thoroughly. Though there are review agents as well, but as an engineer you should review your code so that based on your understanding of systems and knowledge of you know architectures, you should not be exposing something which should not be exposed like with your agents. Like the the core API keys, the configurations, the databases, the authentications and all that stuff. Right? So I think even though AI can do of its best, these are certain things which you should still, you know, check and that's a role of a developer is irreplaceable. You have to do the security audit for everything which is being generated out. You have to be responsible about understanding what the generated code actually means and does. and also does that code really aligns, let's say if you're working in a bigger systems, does your code really aligns with the architectural designs of your system or is it doing something by its own? So I think like No, even if you're not doing something in the bigger systems or bigger code bases, it does the architectural pattern designed by AI will sustain. I think that's all where your developers understanding of architecture and you as a developer become irreplaceable. Right, perfect. I like how you said like you know it's it's an irreplaceable role. But that isn't you know mean that there aren't changes coming. So we've talked about AI and by coding how it has changed software development and what skills and tools developers should work on to stay ahead. On the other hand we also saw like you know where AI and coding tools is still lagging and making mistakes. But the future has a lot of AI in it which is like a known fact to everyone here. So Arpit, what is what is exactly coming? Like what are some trends you're observing in this space? What can developers expect to happen in the next one or two years? The one thing I you know keep saying which is we all will become builders rather than software engineers. So the lines between the ideator, previously there were like PMs and others, product management peoples and others, the thinkers around the designs, the UX people, and the actual implementer like software engineers. Those are changing. Those are blurring. And you all we all are becoming, you know, the builders there. And the agentic development is no more a is no more a optional thing. It's a must thing and it's going full swing ways mainstream. I personally would not have written more than or or augmented more than 5 to 10% of my code in last few months. Everything is reviewed everything is written by agents. Part of it is mostly reviewed by agents too, but complex systems where agents, you know, give output, I'm the one respon like we are the one as an engineer responsible for, you understanding those codes, understanding those systems, understanding those reviews. I also feel like um you know, given let's take an example, right? I have spent couple of my years understanding the apps ecosystem there. Those boundaries are thin. You are expected nowadays you are expected to be learning any new technology with the shortest span of time. You are kind of becoming full stack developer where you should know how these entire systems work because obviously coding in those different and known environments is no more a question of AI can do it, right? Given your details around the thing which you want to do is like perfect. That's essentially which is means like upscaling towards all these agentic assisted coding is not an option. It's mainstream. It's going full stream. You are by default expected to be an AI builder. If you ask someone like hey, are you still writing code? Someone will look down on you and say, hey, who you who write codes in this world of today? It's all agentic coding. And this will evolve and this is going further and further and further in the coming, you know, months or years, right? Okay, great. Thanks for sharing your thoughts on that. Before we move on to the next slide, I see a couple of you have raised your hands. I believe you have some questions to ask or some comments. Um Renuka. So uh please do, you know, use the chat box or the Q&A box, preferably the Q&A box where we would be able to moderate and pick the best questions and share it to our expert here. So please do take just, you know, a minute or so's time to type out the question and let us know. We would definitely try to answer your doubt. Um yes, so moving on Arpit, like while researching for this webinar I read that full stack engineer roles have grown 35% year over year since 2015, which is like you know a phenomenal trend and you know that's continuing even today despite AI and all the changes that we previously discussed. So is full stack development still the way to go in the future? Also is like full stack actually doable for everyone for most people? What what are your thoughts on this? Surely I think full stack is a way. So when when we say full stack it also essentially means for senior engineers about understanding of the entire, you know, systems across as well. Right? Or for the engineers who are in the early stages of the career which also means understanding both the systems front end and back end. So as I said like learning curve has learning curve has gone like the learning time has gone drastically down. So expectations have changed. We are expected you know to have the entire understanding of these systems. The entire full stack understanding because coding in those different systems which you may not be aware of is no more, you know, tough. It's all agents which help you do it. So I think it's surely a no-brainer that full stack is the, you know, path down the line. Okay, perfect. And I think that sets the right, you know, context for us to you know, introduce our program as well. So what I take away from this is that like you know there's not like one specifically one right path but there's always like you know a smarter way to go about this. So for someone who needs to learn full full stack, someone could, you know, need to would want to specialize, someone might want to develop more to knowledge. But the question becomes like how do they even start? And that's the gap that we are trying to close here. So here's the recommended learning pathway to all of you here. Um to become an AI-powered full stack developer, we have a great program that we will be delivering in collaboration with Microsoft. It's a live online interactive program. It spans for 9 months but requires only about 10 hours of your time a week. Um you will have the major Microsoft advantage with classes delivered by a lot of industry experts. There's a very practical learning approach. Here you will be working on more than six projects. You can build your overall like you know software development and AI portfolio and you could also get to improve your coding skills. Um so let's quickly dive into the learning path for becoming a full stack developer. First we'll start with the foundations of front end development. This is where you will learn the basics every developer needs to know to create user-friendly interfaces. Um next we have the GenAI fundamentals for developers. This segment will introduce you to the world of AI and how it can transform your whole development process. Um then you will learn about designing a dynamic front end with React. For those of you who might not be aware, React is a library for building so sophisticated user interfaces. And moving on you'll discover how to harness GenAI from design to code optimization. So this will show you how to utilize AI tools to streamline your coding process. Um and like you know a very very small preview trailer of that is what we have had as a discussion today. We'll also be covering designing and managing databases with MongoDB which would be essential for handling data efficiently. After that you'll focus on developing a reliable back end with Node and Express, two fundamental technologies for server-side development. We'll also explore data structures and algorithms for solving complex programming challenges. Um and as you prepare for your full stack journey, we'll introduce GenAI powered software testing to enhance the quality and efficiency of your um testing process. Um and finally, we'll put all your learning into practice with a capstone project, which is a great opportunity to um apply and showcase all the new skills that you would have acquired. Um each of these modules builds your expertise in prepares you for um the challenges in the work atmosphere out there. Um and the curriculum is designed to ensure that you gain hands-on experience in both front-end and back-end development. Um a unique feature of the program is the integration of AI in software testing and optimization as well. A portion um you know, is what we discussed today in the I mean, we discussed the workflow. Um by the end of the program, you'll have a comprehensive skill set ready to tackle any challenge in the world of uh full stack software development. Um Arpit, it would be great if you can uh come in here. Like, what do you think about these tools, um these framework, platform, skills, anything? Like, how relevant do you think all of these are um for uh software developer today? Got it. I think obviously, even in the assisted way of coding, as I said, learning the basics of these frameworks, tools, is a no-brainer. One should have understanding of couple of these tools. And yeah, so I would say it's a no-brainer to understand these tools. Learn it. AI will help you become 10x if you already know what these tools can do. You can do it 10x faster. Okay, perfect. Uh so, for those of you who are interested, here are the program investment details. Uh for Indian learners, the fee is 53,999 rupees with monthly installments starting at just uh 2,418 rupees. Uh for the learners from the USA, the fee is 1,449 dollars with installments starting at just 125 dollars a month. Uh for our audiences from other parts of the world, you uh apologies for not being able to share um the exact um um you know, program investment in your currency, but um I am sharing the uh program page here. Um you will be able to uh find all the details of the program as well as like, you know, your exact you know, how much the fee is in your exact currency um in this link. Um so, with that, I also going to take a moment to launch a poll. Just give me a second. Yes. Um so, I have now launched a poll to take your interest in enrolling in the program. Uh please do drop in your um answers now. All you have to do is click on yes or no. 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 if you want to have a look at the syllabus much more in detail, um you can scan the QR code here. And in case that you um you know, you're you are um clicking on no, we would love to know like, you know, what is holding you back. So, please do um let us know in the chat what um you know, is holding you back, why you're clicking on no. Okay, okay. So, while we have that poll live, it's going to be live for um a few more minutes. Um Arpit, it would be great if we can address some questions. Sure. Okay, I'm going to read out a couple of questions that have come our way. Um Okay, so this is from Kamalpreet Singh um and this is regarding um the hallucinations that we discussed about um a bit uh before. Um so, how can we find out about these hallucinations by AI, especially for non-coders? Um if there are any issues in the code, how do we verify? Interesting one. I think one way So, I think it's it it requires So, in order to understand the flow, it requires uh a person have basics understanding of coding, right? So, one should know like, in order to understand you know, the you know, difference between wrong and right, you should know what what's wrong and what's right, right? So, one should have a basic understanding of these uh you know, Given these AI are improving at a rapid pace, these hallucinations go away with the details prompt, details skills you add, skills that maybe you add to those prompts or your agents. So, I think there are ways to reduce it. Uh but in order to understand whether the output is fully correct, fully not hallucinated, like for a someone like non-coder, I would say then, you know, try to experience the output. Let's say if you're coding up a web app or something like that, right? Try to experience the output itself and then give them a feedback loop back with the agent saying, "Hey, this is something which I realize is not working. What is an issue? Can you help it out?" That is how how I I would approach it there. As a non-coder. Okay, thank you, Arpit. Um and thank you for that question. I uh hope that was resolved. Um okay, a couple more questions here on the Q&A box. Okay, uh what defines a junior developer today? Like, is the barrier to entry now higher as in like, you know, having scaffold or architecture knowledge or lower like, just having prompting ability? What is required um to be a junior developer today? Interesting one and a good question. I think uh The bar is higher, obviously. First. Right? Even if someone asked, right? Non-coder can do the programming, it means being a coder, being a software engineer, entry-level software engineer, the bar is higher. Maybe prompting in the best way can help you, but you should still I I would always emphasize on having the basics knowledge of computer science. Those things will never go away. Those things are actually will even help you define your prompts, define your your to be much more relevant, much more, you know, useful for the AI agents, too. I understand, you know, given the given this world, how you know, with how much fast pace we are moving, this might be a legit question or thought around people who are, you know, graduating or like, just early into their careers, but I would still say focus on the core of fundamentals. And then use those fundamentals to prompt with AI. And And you'll get the best output there. Okay, great. Thank you for your answer, um Arpit. Let me just There are quite a lot of questions that have come. Just give me a second. Okay, so I'm not sure if you'd be able to uh specifically answer this, but maybe you can give your insights on how to go about this overall. Um this is from Renuka, who says that um you know, learning more about AI prompting requires a lot of uh practice. Um so, um is there a way that you suggest they go about it particularly like, you know, how do they pick up um project um how do they practice on this um as a project? Correct. Uh I would approach in a different angle. I would say learning anything takes practice. Let's take off maths, right? Or any any of your any of your favorite subject. There are just few patterns which you have to understand to cover almost 80 and 90% of the ground. So, I think this is something one should learn. And I think there are maybe courses around those which you should understand how to which will give you the idea of how to best prompt. Right? Is prompting the right way even, right? Beyond prompting, there are things which I talked about, skills and others, right? Things are evolving there. Is prompting the right way? Is Is skills the right way? Do you even need prompts? Do you even need skills? Right? So, these are the set of questions one should understand first. And then follow certain best practices which you can learn from either courses or or Google or or somewhere. Right? And I think that's how you'll you'll cover almost 90 and 95% of the grounds of best practices in prompting. So, that's how I would approach that. Great. Um so, with that, we've gotten quite a lot of interest in uh enrolling in the program. I'm going to end the poll here. Um and I know that quite a lot of you might be interested in the attendance certificate as well. Um, so here is a poll that I have I'm launching for this. Uh, please ensure to enter your full name to receive the certificate of participation. We don't need your email or any other details, just the full name on like you know how you want it to be on the certificate, just enter that and um submit uh the poll. We will be able to generate the certificate and share it with you in the next um 24 working hours. I think I think we can take like this last question, maybe. Um we'll see how long people are taking time to answer the poll. Um there's someone who has asked, are there any like you know tools that can be used or integrated in VS Code um specifically to help in vibe coding bug fixing? If you have any tool recommendations, they are free tools and I think Right. Have a good one. Tools is something like again, like don't rely on any one of these agents. Obviously, one one let's say XYZ agent works best, you know, for certain scenarios if they you know, if if compared to ABC, right? So I think there are common patterns emerging within all these tools. And more or less, I I think like all these tools, companies, Claude, Gemini, ChatGPT, whatnot, will eventually end up becoming almost the same. Uh they'll have more or less the same capability for the user. Some for the some time maybe working best better than other, but I would say don't emphasize on one tool, rather emphasize on how these tools really work. What works best across the tools. Prompting will always work best. The one of the the question was specifically around testing, maybe? If I'm not wrong. For testing bug fixing in testings, right? Got it. So I think what I in general in day-to-day uses like agents with skills. Given the bug with the details, and it go find out code bases, find out the exact patterns, find out the flaws, and I obviously, you know, uh give a feedback loop in their thinking process to make it more accurate. So I think one of one a few things that I have to quote like one is like giving the feedback loop to the agent back in their during their thinking process, not in the execution process really works. Your bugs should have best of details or most of details captured in there, these steps to repro those bugs. I think that's where these autonomous Once you Once you have all these details, that's where these autonomous agents will go really and figure out your code base, understand from the bug, figure out your code base, try to fix it, provide you a thinking process. You providing back the feedback loop, they'll improve and actually eventually do it by by by its own. So I think that's how one should approach this. Okay, perfect. Thank you so much uh Arpit. Uh with that, I am also ending the certificate poll. Um and uh to all the participants, thank you very very much for spending the last 1 hour with us. Uh we hope that you'll be able to apply the learnings from this session um as well as what you'll get from our AI-powered full stack developer program to get ahead in your career. Have a great day. Thank you. Arpit, once again, thank you so much for your guidance and sharing your experience and your knowledge. This conversation was very insightful for me as well. Uh it's been a pleasure hosting you. Do you have have any final remarks to share? Any like quick advice for our learners? Become a builder. Ship fast. That's the only thing this world needs. Become a builder. You don't You should not, you know, restrict yourself to, "Hey, I just know this. I can do this." If you don't know, learn because the learning curve has gone down. Become a builder. And and likewise, Anna, thank you for having me here. I think I really enjoyed the entire discussion with you and the, you know, participants. Thanks a lot. Perfect. Thank you so much. Uh with that, we're wrapping up the session. We know that we were not able to answer all the questions, but uh we just tried to ensure that we try um you know, we cover as many of your questions as possible. So if there are like multiple questions from one person, we take up just that and we try to ensure that we cover as many um you know, people's doubts and queries as possible. So I hope that everyone understands that. Uh but if um you know, if there is anything that we can support you with, if there's anything that we were not able to address, or if you have any feedback to share, please do write to us at [email protected]. We'd love to hear from you. With that, I am ending the session. Thank you um all for being here.