AI/ML engineer. Computer vision, language models, systems that run in real time. Building from the math up — implementing architectures, not just calling APIs.
| Repo | What it does |
|---|---|
| AttentionIsAllYouNeed | Transformer from scratch in PyTorch — multi-head attention, positional encoding, full encoder-decoder |
| RealTimeASLTranslator | Real-time ASL classification via CNN + MediaPipe hand landmark pipeline over live webcam |
| TribunalAI | GPT-4o, Claude Sonnet, and Gemini Flash debate independently, critique each other, then synthesize consensus |
| Desktop-Nanny | YOLOv8 screen monitor — detects objects in real time and triggers configurable alerts or blocks |
| LexonAI | Source-transparent AI answers with 3D knowledge graph visualization — MadHacks 2025 finalist |
Python · PyTorch · OpenCV · MediaPipe · YOLOv8 · FastAPI
TypeScript · Next.js · OpenAI · Anthropic · Gemini
Training models from scratch, working through transformer internals and scaling to more architectures. Next: production model deployment.


