AI Engineer & Data Scientist β I build LLM agents, MLOps pipelines, and chat-first tools that ship.
- π₯ Stash β a chat-first "save anything" bot: forward a link or note to Telegram, an LLM auto-tags it, and semantic search finds it later. Serverless on Cloudflare Workers (TypeScript, D1, Workers AI). Private beta.
- π Enterprise AI integration β a Redis-based bridge connecting cloud AI apps (document intelligence, order assistants, chat) to on-prem ERP systems, plus an AI document summarizer for finance teams (Azure OpenAI).
- π¬ ReelDrop β open-source brand-aware video engine: a sentence + a brand kit in, finished videos, shorts, and decks out.
Languages β Python Β· TypeScript Β· SQL
AI / ML β LangChain Β· LangGraph Β· Agents Β· RAG Β· Azure OpenAI Β· TensorFlow Β· scikit-learn Β· MLflow Β· DVC
Platform β Cloudflare Workers Β· Azure Β· Docker Β· Redis Β· FastAPI Β· Git/GitHub Actions
| Project | What it shows |
|---|---|
| Agentic LangGraph Chatbot | Multi-step agent orchestration with LangGraph |
| ML Pipeline with DVC | Reproducible, versioned ML pipelines |
| MLflow Experiments | Experiment tracking & model registry workflows |
| US Visa Approval Prediction | End-to-end ML system with deployment |
| Pneumonia Detection Web App | Deep learning served as a usable web app |
| Stock Trend Prediction | Time-series forecasting with deep learning |
From notebooks to production β I like taking ideas all the way to something people can use.
