AI Engineer with a production reliability background. I build LLM pipelines, retrieval systems, and internal AI tooling that engineers actually use. I bring production support instincts to AI infrastructure. Measure first, calibrate second, ship what survives real workflows.
- 10 years in Linux infrastructure and technical support.
- 10% AI ticket deflection in production workflows.
- 3x Top Engineer of the Quarter, 1.0 CSAT in 2026.
Based in Lins, São Paulo, Brazil. Remote since 2020.
Selected Work
dubweave
Fully local AI dubbing pipeline. Download, transcribe, translate, and dub YouTube videos, with nothing leaving your machine. No cloud APIs, no subscriptions.
View Project
Zendesk Real-Time Notifier
Chrome extension for real-time Zendesk ticket monitoring. Built to solve missed tickets while context-switching between browser tabs during high-volume support shifts.
View Project
mcp-plesk-dev-docs
MCP server for semantic search over the Plesk documentation corpus. Vector embeddings, cross encoder reranking, and RAG evaluation.
View Project
What I Work On
- Retrieval systems and semantic search.
- LLM pipelines with evaluation, calibration, and quality gates.
- Internal AI tooling for support and operations teams.
- Automation across Linux, APIs, and developer workflows.
Background
I build AI systems with a production reliability mindset shaped by 10 years of Linux infrastructure and support engineering. I focus on real workflows, measurable outcomes, and tooling that reduces operational load.