A curated selection of projects spanning AI agents, retrieval-augmented generation, data engineering, and MLOps.
9 projectsReal-time data visualization platform with machine learning insights for enterprise clients.
Retrieval-augmented generation system enabling semantic search across millions of enterprise documents.
Built a retrieval-augmented generation system on Azure AI Search that extracts page-number citations at query time — no re-indexing required — achieving 94% citation accuracy across enterprise document collections.
Built an agentic document validation system for Scoop using LangGraph and GPT-4o, achieving 97% validation accuracy through a self-critique loop and parallel batch processing.
Replaced daily S3 batch ingestion with a near-real-time AWS streaming pipeline, cutting end-to-end latency from 8 hours to under 4 minutes at 50M events/day.
LangGraph-based agent system coordinating specialized AI workers for complex task automation.
Scalable data pipeline processing streaming events from multiple sources into a unified warehouse.
End-to-end MLOps platform for versioning, deploying, and monitoring ML models at scale.
Conversational AI assistant with RAG-powered context retrieval for internal documentation.