PROJECT / 01
NYAYA - AI-Driven Research Engine for Law
A legal research engine that combines fine-tuned language models with retrieval-augmented generation to surface case precedents and structure findings for commercial-court workflows.
- Python
- Fine-tuned LLMs
- RAG
- FAISS
- multi-qa-mpnet-base-dot-v1
SYSTEM BUILD
What went into it
- 01
Built ingestion pipelines for India Code, legislative.gov.in, and eGazette documents.
- 02
Implemented section-level chunking, custom metadata, quality remediation, retrieval, and contextual reasoning.
- 03
Generated structured summaries and recommendations from retrieved historical legal material.
MEASURED OUTPUT
What changed
Produced a purpose-built legal corpus of approximately 5,752 searchable chunks spanning major Indian statutes.
Created an end-to-end research workflow that connects precedent retrieval with decision-support output.