omkar thakur

I’m a final-year B.Tech student in AI & Data Science (8th semester completed) with hands-on experience turning optimization and machine learning into usable software. I focus on clear problem framing, fast iteration, and building systems that can run in real-world workflows.

In my Data Operations internship, I performed exploratory data analysis and built predictive reports on 500+ investment records, improving funding-stage classification accuracy by 15%. I also developed and deployed ML classification models with 87% precision to support data-driven investment decisions across teams.

My project work centers on optimization democratization and production readiness. I built Optopus, a supply chain optimization platform with a Flutter desktop client and a Python FastAPI/MILP backend, including an interactive Map View using OpenStreetMap/OSRM and a Supabase-backed time-bucket synchronization system for consistent period-based planning.

I’ve also created a Visual MILP modeling platform with drag-and-drop MILP building and GNN-based solver selection across Gurobi and CPLEX, reducing solution time by 30%. Alongside that, I developed a Cython-based MathRice library achieving a 5x speedup for matrix operations and published research on making MILP accessible to non-experts.