Practical Generative AI for Coders

LIVE

24 Hours

₹ 22,000 per month.

₹ 24,000 1-on-1 Class

Just Pay ₹11,000 to Register

- OR -

Course offered by Sanjay Bhatikar

0 review

overview batches reviews

Six weeks. Six notebooks.
One capstone.

Week 01: Tokens, Transformers & Your First LLM Call

Build a BPE tokenizer from scratch, load GPT-2, and experiment with decoding strategies live.

#BPE tokenization #HuggingFace basics #Temperature & sampling #Model anatomy

CPU only

Week 02: Prompting as Programming

Build structured output pipelines and an eval harness. Write a chatbot with conversation memory.

#Few-shot & CoT #Structured JSON #Eval harness #Multi-turn chat

API only

Week 03: Fine-Tuning with LoRA / QLoRA

Fine-tune a 1B model in ~45 minutes on Kaggle's free T4. Compare against the base model on held-out data.

LoRA intuition 4-bit quantization Instruction datasets Loss curves

T4 GPU

Week 04: RAG — Retrieval-Augmented Generation

Build a vector store in NumPy first. Then build a full Q&A system over documents with ChromaDB and hybrid search.

#Embeddings #Chunking strategies #ChromaDB RAG evaluation

CPU only

Week 05: Agents & Tool Use

Implement the ReAct loop from scratch, then use Anthropic's native tool-use API. Build a multi-tool research agent.

#ReAct pattern #Function calling #Error recovery #Agent evals

API only

Week 06: Multimodal Models + Capstone

Vision-language models, multimodal RAG, and deploy your capstone app to HuggingFace Spaces with Gradio.

#Vision LLMs #Multimodal RAG #Gradio deployment #Capstone project

About the Trainer

Avg Rating

Reviews

Students

Courses

Sanjay Bhatikar

Master of Engineering - Master of Technology (M.E./M.Tech.) from University of Colorado Boulder in 1999 and Doctor of Philosophy (Ph.D.) from University of Colorado at Boulder in 2003

20 Years of Experience

I hold a PhD in AI/ML from the University of Colorado Boulder and am the co-founder & CTO of an AI-driven FinTech firm transforming unstructured data into quantitative market insights.

With 20+ years of hands-on AI/ML experience across global giants like GE Global Research, Monsanto, and Japan Railways, I bridge the gap between academic theory and real-world production systems. My work spans deep learning (CNNs, neural networks), OCR, and modern Generative AI architecture.

What You Will Learn in "Practical Gen AI For Coders":
Production-Grade Gen AI: Moving beyond basic prompt engineering to building real LLM applications, custom RAG pipelines, and agentic workflows.
Industry Best Practices: Architecture designs, tool selection, and code optimization used in actual corporate and startup environments.
Hands-On Code: Write Python code from day one—no fluff, just practical, deployable projects.
Whether you're a software developer looking to transition into AI or an experienced engineer wanting to master Gen AI architecture, I focus on making complex concepts simple, practical, and immediately actionable.