About

I'm a Lead ML Engineer based in São Paulo, Brazil, with 12+ years of software engineering experience. I currently architect real-time computer vision and LLM inference pipelines for live video production at Monks (Media.Monks).

My work sits at the intersection of ML infrastructure, autonomous agents, and quantitative systems. I design and deploy multi-model ML systems across GPU clusters — orchestrating Transformer-based models, Vision-Language Models, and LLMs on NVIDIA L40S hardware via Kubernetes.

Before ML, I spent a decade in enterprise distributed systems across financial services and government — Banco do Brasil (South America's largest bank), nationwide tax systems for Angola's Ministry of Finance (IMF compliance), and insurance platforms handling 500K+ daily transactions.

On the side, I run Curupira, an open-source quant research lab where an autonomous AI agent conducts its own trading research with full transparency — including publishing failed strategies.

Career Timeline

Lead ML Engineer

Monks (Media.Monks)

Real-time ML platform for autonomous sports video production. 6 ML pipelines, 7+ models across 4x L40S GPUs, LLM-powered autonomous video director.

Senior Backend Engineer

Mapfre

Cloud migration for insurance pricing. AWS serverless microservices handling 500K+ daily transactions.

Tech Lead

Tis Tech, Ministry of Finance — Angola

Led multinational team building nationwide tax collection systems for IMF compliance. 99.99% uptime, mentored 10+ developers.

Senior Software Engineer

Mirante Tecnologia

Production support for Brazil's largest credit union. Database optimization and REST microservices.

Java Developer

LinkData S.A.

Government asset management with QR code/RFID tracking across agencies.

Junior Java Developer

Banco do Brasil

Offline banking solution for remote Amazon communities with satellite data transmission.

Education

Bachelor's in Computer Science — UniCEUB, Brasília, Brazil (2017)

Core Skills

ML / Computer Vision

PyTorchObject DetectionSegmentation Pose EstimationVLMsvLLMHuggingFace

ML Infrastructure

NVIDIA NIMCUDAGPU Time-Slicing MilvusModel Serving

Languages

PythonJavaSQL JavaScriptC#/.NET

Cloud & Infra

KubernetesDockerAWS KafkaPostgreSQLMongoDBElasticsearch