AI Engineer · Systems-minded
I work on AI systems that are expected to run in production.
My focus is on correctness, cost, observability, and long-term maintainability.
- Design systems first, models second
- Prefer simple, explicit architectures over clever abstractions
- Optimize for debuggability, cost control, and failure modes
- Treat AI components as infrastructure, not experiments
I’m still learning every day — especially where AI meets real-world constraints.
AI & LLM Systems
- Retrieval-Augmented Generation (text & multimodal)
- Agentic workflows with deterministic execution
- Model routing, evaluation, and cost-aware inference
- Open-source + proprietary model tradeoffs
Backend & Platform
- Python, FastAPI, langgraph
- PostgreSQL, Redis
- Docker, Kubernetes
- AWS / Azure
- Structured logging, metrics, traces
Product Engineering
- Building internal AI platforms teams can trust
- Clear APIs, versioning, and operational discipline
- Frontends only when they serve the system (Next.js)
| Project | What it explores |
|---|---|
| RAG Pipeline Optimizer | Comparing RAG configurations to understand accuracy vs cost tradeoffs |
| Cost-Control Model Router | Routing prompts to the least expensive capable model |
| Multimodal Video RAG | Searching visual moments in long-form video using VLMs |
These projects are built to understand tradeoffs, not to chase benchmarks.
I value clarity over cleverness, and systems that age well.