At Intuitive.ai, I worked on an internal engineering accelerator, building a RAG pipeline over internal documentation with Qdrant vector search and Cohere reranking. I also developed a multi-agent system with Google ADK to automate a multi-step internal workflow.
I built async FastAPI endpoints for the pipeline with JWT authentication, streaming responses, and request tracing, and managed the Spring backend for frontend. I also worked with React.js and Tailwind CSS on user interfaces.
On the group project CodeBase Brain, I built parts of a RAG-based codebase Q&A system that ingests GitHub repositories through a six-phase pipeline. The project used Tree-sitter parsing, Gemini on Vertex AI, Qdrant, and async indexing with Redis/RQ workers and PostgreSQL.
I also built Vidya Study AI Companion, a full-stack learning platform where teachers upload study materials to AWS S3 and students access them and generate AI-powered summaries. The backend used both .NET Web API and Spring Boot.
