RAG-Based Study Assistant
RAG pipeline over lecture transcripts with semantic chunking and context-aware retrieval for personalized study guidance at scale.
I design and ship production AI systems: RAG, Agentic AI, and LLM-powered services that real users depend on every day.
Production AI systems in EdTech and E-Commerce: RAG pipelines, agentic workflows, and LLM-powered services running for real users.
Production RAG systems serving real learners: 60% faster content discovery, 15% higher course completion.
Text-to-SQL analytics enabling 50+ non-technical users to query databases in plain English.
Agentic AI systems automating manual workflows from 200 hours per week to 10 minutes.
RAG pipeline over lecture transcripts with semantic chunking and context-aware retrieval for personalized study guidance at scale.
50+ non-technical users querying production databases in natural language. Built on LangChain, FastAPI, and Streamlit.
Multi-modal pipeline: speech-to-text + LLM rubric evaluation to generate performance reports from session recordings.
RAG classification pipeline for ECCN trade compliance. Retrieval accuracy from 65% to 86.5% at scale.
Alongside engineering, I occasionally run sessions on RAG, Agentic AI, and LLM systems, teaching from systems I've actually built and shipped.
If you're building production AI (RAG, Agentic AI, or LLM-powered systems) and need an engineer who's shipped this before, reach out. I respond within 24 hours.