Tech Stack: React, JavaScript, FastAPI, Python, Groq, Ollama, Qdrant, Supabase, BM25, Reranking, Vercel, Render

Legal contract review is prohibitively expensive for individuals and small businesses, with enterprise software solutions costing thousands of dollars — leaving most users without access to reliable legal risk analysis.

Architected and developed a two-stage AI-powered contract review platform from the ground up. Designed a document ingestion and chunking pipeline to handle legal agreements of varying complexity. Implemented a hybrid retrieval system combining vector search (Qdrant) and BM25 keyword search with reranking to surface the most contextually relevant clauses. Integrated Groq-hosted models for fast initial analysis and Ollama for deeper local verification, enabling a dual-confidence review workflow. Built an evaluation-driven scoring system to quantify risk and identify missing contractual protections, exposed via a FastAPI backend and consumed by a React frontend.

Delivered a fully functional, production-deployed legal intelligence platform that replicates capabilities found in enterprise legal software — built entirely at zero cost. Demonstrated proficiency in full-stack product design, RAG pipeline architecture, hybrid search systems, and legal-tech workflow engineering.