Build a production ready RAG system from scratch with Raj Abhijit Dandekar
Raj Abhijit Dandekar
Build a production ready RAG system from scratch
Digital Product
About this product
The workshop is titled “Build a production ready RAG system from scratch”.
Instructor: Dr. Raj Dandekar, MIT PhD
Timings: 8am IST on Saturday, Sunday (23-24 August)
Here is what we will cover:
(1) Build an end to end RAG pipeline from scratch
(2) Compare the different data ingestion approaches:
- PyMuPDF
- Tesseract OCR
- Docling
(2) Code and compare 5 chunking strategies:
- Fixed chunking
- Recursive chunking
- Structural chunking
- Semantic chunking
- LLM chunking
(3) Compare 3 types of embedding strategies:
- PyTorch embeddings
- Vector Stores (Pinecone, Weaviate, Chroma)
- Database integrated vector stores like Postgres + pgvector
(3) Run the pipeline locally using open source LLMs
(4) Implement a RAG Evaluation pipeline using RAGAS:
- context precision,
- context recall,
- answer relevancy,
- faithfulness
(5) Deploy the RAG pipeline using Supabase + Postgres + PGVector
(6) How to design multimodal RAG pipelines
(7) Using Lovable to make beautiful frontend
Check attached video to see what we will be building!
P.S: The 3 hour workshop may extend to 4 or 5 hours!
What are people saying
The session was extremely insightful and well-structured. It clearly reflected the depth and seriousness behind the initiative. It’s rare to find such a thoughtfully designed program in the AI/ML research space, and I appreciate the effort put into creating a platform for committed learners and researchers.
Nishant Kashyap
Jul 2025
You approach towards machine learning and AI is mathematically very insightful which helps to know how things are working in background which is very necessary to build good foundation in any subject.
Saurabh Sharma
Jul 2025
It was insightful and worth my time. I appreciated how clearly you explained the strategies for developing a strong research direction
Anonymous
Jul 2025
₹3,000