- B.Tech in Information Technology @ VJTI, Mumbai (CGPA: 8.05/10.0) — Class of 2028
- Passionate about AI/LLM systems, especially RAG pipelines and multi-agent architectures
- Full-stack developer comfortable across React/Next.js frontends and Node.js/FastAPI backends
- Based in Mumbai, India
- Reach me at vedantshimpi2006@gmail.com
Languages
Frontend
Backend / APIs
Databases
AI / LLM
Tools
🔍 RagNexus — LangChain • Qdrant • Python • BM25 • Transformers
A modular, production-grade RAG pipeline for enterprise document Q&A at scale. Combines dense vector embeddings (Qdrant) with BM25 sparse retrieval through weighted fusion, layered with HyDE query expansion and a cross-encoder (MiniLM) reranker to boost precision. Evaluated using RAGAs metrics (faithfulness, answer relevancy, context recall). Hardened with JWT auth, per-tenant data isolation, and rate limiting for multi-tenant use.
🧠 ResearchMind — FastAPI • LangChain • Python • BeautifulSoup • React • Groq
A local multi-agent research assistant orchestrating a 4-phase pipeline (search → scrape → draft → critique) using LangChain agents. Features a writer–critic chain architecture where one agent drafts structured reports and another scores/critiques them against a fixed rubric — all exposed via a FastAPI + React app.
🧪 LabTrack — React.js • Node.js • Express.js • Supabase • Socket.IO • Tailwind CSS
A full-stack lab asset & complaint management system with role-based access control, real-time Kanban complaint tracking via Socket.IO, AI-assisted complaint prioritization with SLA monitoring, and analytics dashboards with CSV/Excel/PDF export.
⭐️ Always open to collaborating on AI, RAG, and full-stack projects — feel free to reach out!