| Area | What I do |
|---|---|
| LLM Applications | RAG systems, AI agents, document intelligence, structured extraction |
| ML Pipelines | End-to-end training, fine-tuning (LoRA / QLoRA), evaluation harnesses |
| Full-stack AI | Next.js frontends backed by LLM APIs, real-time AI UIs |
| MLOps | Model serving, observability, CI/CD for ML, experiment tracking |
- Building AI-powered B2B payment orchestration — invoice extraction via Gemini 2.5 Flash
- Exploring agentic workflows with tool-use and multi-step reasoning
- Shipping production RAG systems with hybrid search + re-ranking
- Reading: Designing Machine Learning Systems — Chip Huyen
🌸 "The best model is the one that ships." 🌸