Case Study

DecoverAI: AI-Powered Legal Intelligence

How I helped build the GenAI platform behind a US legal AI company that raised $2M+ in seed funding

Yash Soni · About 13 months, 2024

The Problem

DecoverAI set out to give US law firms semantic search over large sets of legal documents.

The Solution

DecoverAI is an AI-powered legal document management and semantic search platform. It uses vector storage and a RAG pipeline so lawyers can query large document sets in natural language.

Architecture & Tech Stack

• Frontend: React
• Backend: Node.js, Python, TypeScript
• AI/ML: OpenAI, RAG, HuggingFace
• Vector DB: Pinecone
• Database: MongoDB
• Infrastructure: Docker Compose, AWS EKS

Key Features

• Semantic Document Search: Natural language queries over legal documents using vector storage and semantic retrieval
• RAG Pipeline: Retrieval-augmented answers over large legal document sets
• Production Deployment: Staging and production environments on AWS EKS, orchestrated with Docker Compose

Results & Impact

• $2M+ seed funding from Bay Area VCs
• 13 months of development
• 10+ US law firms in production

My Role

I was a contractor on this engagement for about 13 months. I helped build the infrastructure and the GenAI platform, helped lead a team, built critical pieces, and tested and deployed them.

“Yash is one of the hardest-working and meticulous engineers I have encountered. We worked closely together on DecoverAI, where he helped build the infrastructure and Generative AI platform that is now deployed in large law firms throughout the United States. He helped lead a team, build critical pieces, and test and deploy them. We eventually raised 2M+ in funding with the help of his efforts. I would certainly recommend him for any future projects.” Ravi Tandon, CEO & Co-Founder, DecoverAI

Last updated: 2026-10-11