I build production-grade data pipelines and AI/LLM systems - from GPU inference and RAG to fault-tolerant ETL on Kubernetes. Grounded in a strong applied mathematics background.
Data & AI Engineer based in São Paulo, Brazil, with 4+ years designing scalable data architectures, deploying GPU-based inference systems, and shipping LLM-powered applications for real-world use cases. Currently a Data Engineering, Management & Governance Analyst at Accenture. Strong mathematical foundation (BASc, USP) and a track record of turning heterogeneous, messy sources into dependable, high-impact datasets.
A self-directed project to push my AI-platform skills - agents, RAG, MCP and observability on AWS.
Backend built on a hexagonal architecture (ports & adapters), exposing platform capabilities to AI agents and grounded on semantic search.
MCP tools, RAG (Titan V2 + pgvector) and LangGraph orchestration, instrumented end to end.
vLLM system that analyzes large-scale legal documents and delivers automated decision summaries.
Replaced a paid embedding solution with an open-source GPU stack - better accuracy, lower cost.
Fault-tolerant, containerized workloads with Airflow on Kubernetes - 100+ tasks per run.
Interested in my work, or want to trade notes on data & AI engineering? I'm always happy to connect.