On NexusMind, I implemented hybrid-retrieval RAG with cross-encoder reranking and per-user scoped retrieval. I evaluated retrieval and answer quality with RAGAS and streamed source-cited answers through a FastAPI backend.
For CareerPilot, I built a LangGraph job-search agent that searches, ranks, and explains listings in a streaming chat. I added a Groq, Gemini, and Mistral fallback chain and checkpointed conversation state in PostgreSQL.
At WorldQuant, I conducted quantitative research on systematic trading and alpha discovery using statistical modeling and large financial datasets. I designed and evaluated signals and developed backtesting frameworks to assess performance, robustness, and risk-adjusted returns.
I'm a B.Tech Computer Science and Engineering student at KIET, focused on LLM applications and Python backends. I also ranked among the Top 45 teams in Hack the Throne 2026 and qualified for Round 2 of the ForkIT Challenge organized by IIIT Delhi.
