Multimodal Agent Evaluation Sandbox
Python, PyTorch, Gemini API, FastAPI, React, PostgreSQL, Docker
A benchmark environment for evaluating software agents across text, image, code, and tabular reasoning tasks.
Merchant Inventory Assistant
Python, TypeScript, LLM APIs, scikit-learn, React, Node/Express
An LLM-powered inventory assistant that routes merchant questions into structured workflow tools and database queries.
3D Data Sandbox
Electron, Node.js, Three.js, PapaParse
A desktop app for offline 3D visualization of CSV and JSON datasets, with clustering, outlier detection, and interactive highlighting.