Apna College

Great fun to be back in Bengaluru✨ Got invited for an engaging conversation with OpenAI CTO to improve Codex for software engineers. Tech has already transformed a lot with AI & many major organizations are working to enhance developers’ experience - not just with models but the entire ecosystem.

An engineer asked me in a recent AI session, why do we need fine-tuning when we already have RAG? Why you should care? Because these are real world AI concepts being used in production systems today. If you are new to RAG, let’s first understand it. Think of it like an open book exam for LLMs. Usually LLMs get a query & based on that model generates a response. With RAG, LLM gets access to a trusted data source (often a Database) to fact check & improve its response. That’s why it’s called Retrieval Augmented Generation (RAG). 🔹 Retrieval - Instead of relying only on what a model learned during training, it retrieves the most relevant info from your own data (docs, PDFs, databases etc.) [fyi - this is called context] 🔹 Augment - In RAG pipeline we augment the prompt with that context 🔹 Generation - The ans that LLM now generates is grounded in real, up-to-date facts Why do we need fine-tuning if RAG gives relevant data to LLM? Because both solve different problems. 🔸 RAG - fetches facts from our internal data, so it suits changing knowledge, private data & ans that need citations. 🔸 Fine-tuning - further trains a pre-trained model on examples to change how it behaves (format, tone, specialized tasks), but it can't reliably teach new facts or keep them current. In short, RAG gives the model new knowledge(context) at answer time (without changing its behaviour), while fine-tuning changes how the model itself behaves. There are definitely more nuances to RAG and I have explained them in a dedicated session on Apna College Youtube✨ (Link in comments)

Thank you, Infosys , for this opportunity! ❤️ I’m extremely happy and grateful to share that I’ve been selected for the Digital Specialist Engineer (DSE) role at Infosys! 🎉 What makes this moment even more special is the date: 18th September 2023 → 18th September 2026❤️. I still remember walking into college on 18th September 2023, knowing almost nothing about coding, programming, or DSA. In my 3rd semester, we had Java taught by Arun K H Sir. The way he taught made me genuinely interested in programming, and slowly, I fell in love with Java. ❤️ A special thank you to Shradha Khapra too. I learned so much from her Alpha Batch, especially programming and DSA. Her teaching became an important part of my learning journey. Then came the countless nights of DSA — even when I was tired, demotivated, or had no confidence, I still tried to solve at least one problem. There were failures, rejections, confusion and many moments of self-doubt. But I kept going. And today, I truly believe — hard work pays off. Grateful to my family, teachers, mentors, friends, Arun Sir, Shraddha Ma’am, and everyone who believed in me. ❤️ To the person who entered college three years ago without knowing where this journey would lead: We made it. 🥹❤️ But this is not the end. It’s just the start. 🚀 #Infosys #DSE #DigitalSpecialistEngineer #Java #DSA #Placement #Gratitude #HardWork #NewBeginnings

🚀 Course Completed | Prime 2.0 (AI/ML) Happy to share that I’ve successfully completed Prime 2.0 (AI/ML) from Apna College. This journey helped me strengthen my understanding of Artificial Intelligence & Machine Learning and build a stronger foundation to apply these concepts in practical scenarios. A big thank you to Shradha Khapra and the Apna College team for creating and delivering such a structured learning experience. 🙌 But completing the course is just the beginning. 🔨 Next: Taking the learning into the real world by working on practical, real-world AI/ML projects and focusing on building things that solve actual problems. Learn → Build → Experiment → Improve. #AI #MachineLearning #ArtificialIntelligence #AIML #ApnaCollege #ShradhaKhapra #LearningJourney #Tech #SoftwareEngineering #Projects #BuildInPublic #ContinuousLearning

There's this specific Wednesday I keep coming back to. Cisco interview in the morning, and Confluent OA was right after. Cisco was the one I'd wanted since I started this whole thing but I got rejected. A few hours later, I was sitting for the Confluent OA. I didn't really get time to feel bad about it before the OA started, and looking back that was probably the best thing that could've happened. I just had to sit down and solve the problems in front of me with whatever composure I could fake until it stopped being fake. The thing I keep telling people who ask me about this. Coming from Civil Engineering, plenty of people, including seniors, will tell you, directly or not, that this isn't really for you. I believed I could do it before I had any real proof that I could. Watching a few seniors from the same branch actually make it helped, but the belief came first, and that was the most important part. If there's one thing from that whole stretch worth passing on is to keep showing up for whatever's next, hackathon, OA, interview, doesn't matter. Every rejection taught me something the last one didn't. Your branch or your CGPA is a starting point, not a ceiling. None of this happens without AMAN DHATTARWAL sir. I've followed his content and the way he thinks about learning since class 11, long before any of this was a real plan for me. Getting to actually sit and talk about my own journey with Shradha Khapra ma'am, and being featured in Apna College, felt like a full circle moment I didn't expect to get. Full conversation here : https://lnkd.in/dvXqFyX7 #SoftwareEngineering #CareerJourney #Confluent

AI is undoubtedly part of every software engineer’s work right now. Companies these days require engineers who have a programming aptitude, can logically build systems & leverage different AI Tools to be more productive. AI is not here to think for us but it’s us that have to learn how to THINK WITH AI. And to help out exactly with that vision we recently put out a free “Complete Masterclass on Codex” on Apna College Youtube in collaboration with OpenAI India to help coders understand how to use the tool to become a 10x programmer. The link for the masterclass is in the comments. Hope this one helps you immensely in your programming journey, especially if you’re still figuring out how AI fits into the way you code. 🚀