Intelligent CCPA Compliance Analyzer
A high-performance RAG system designed to evaluate business practices against the California Consumer Privacy Act (CCPA) using AI-driven reasoning.
Project Overview
Developed for OpenHack 2026 (IISC Bangalore), this project provides a structured, legalistic evaluation of business practices against the CCPA statute. It ensures that AI responses are grounded in actual legal text rather than general knowledge.
The system utilizes a Retrieval-Augmented Generation (RAG) architecture to parse the CCPA statute into distinct subsections, allowing for precise semantic retrieval and structured LLM reasoning.
Key Challenges
- Implementing a robust RAG pipeline to ensure high accuracy in legal text retrieval using ChromaDB and
bge-small. - Designing structured system prompts to guide the LLM (Gemini 2.5 Pro) in generating formal compliance verdicts.
- Optimizing a unified Docker deployment containing both the FastAPI backend and React frontend behind an Nginx reverse proxy.
- Managing real-time feedback for users through Server-Sent Events (SSE).
Outcome
Successfully built a production-ready compliance tool that provides clear, justifiable verdicts for complex privacy scenarios. The project demonstrated the effective integration of vector databases and generative AI for specialized legal analysis, earning recognition at the OpenHack 2026 competition.