AI Engineer · agents, retrieval and the infrastructure that keeps them honest
I build generative AI systems that hold up in production. Backend engineer by training (Java/Spring), now working in Python/FastAPI on LLM orchestration, retrieval and evaluation.
- Agentic systems — multi-agent flows with LangGraph, LangChain and Google ADK; tools, memory and policies exposed over MCP and A2A.
- Retrieval that survives review — chunking, pgvector/FAISS, re-ranking and continuous evaluation with RAGAS. Answers that cite their source, or admit they can't decide.
- APIs and pipelines — FastAPI services with auth, rate limiting, tests and CI/CD; event-driven flows on Kafka.
- Shipping and watching — Docker and Kubernetes on GCP/Vertex AI and AWS, traced with LangSmith and OpenTelemetry.
| 🥇 Winner | Summer Hackathon 2025 — MongoDB & Telefónica. Multi-agent + RAG system to detect illegal ticket resale. |
| 🥇 Winner | TheGameIsHackathON 2024 — CaixaBank Tech & NUWE. Backend track, 1,700+ participants. |
| 🛰️ Finalist | NASA Space Apps Challenge 2025. RAG and an LLM-built knowledge graph over space-biology papers. |
Pinned repositories below · full background on LinkedIn

