At AISPRY, I cleaned, transformed, and analyzed more than 50,000 records, improving data quality by about 20% through validation and outlier handling.
I analyzed enterprise datasets with hypothesis testing and correlation analysis, and visualized pricing trends for pharmaceutical companies. I also authored data-driven reports that translated findings into recommendations for stakeholders.
I built RepoMind, a retrieval platform that indexes codebases into a searchable knowledge base. Its pipeline creates vector embeddings and stores them in Supabase/pgvector, while FastAPI serves a multi-agent layer coordinating LangGraph with Groq and Ollama.
I also built SmartSort AI, a desktop application that organizes files using semantic content analysis. It combines embedding-based search with duplicate detection and storage analytics.

