|
I am a Computational Engineer graduated from UFJF (Universidade Federal de Juiz de Fora) and a passionate Python Developer. My focus is on Data Science, Machine Learning, and Computer Vision. I am dedicated to building robust intelligent systems, from training complex models to deploying scalable MLOps pipelines in critical production environments. |
|
|
|
|
|
- Machine Learning & Deep Learning: Developed and fine-tuned models (ResNet18, Swin Transformer) for detection, classification, and tracking in critical sectors (onshore, offshore, railway). Implemented Explainable AI (Saliency maps, Integrated Gradients) and custom algorithms like ICP for trajectory estimation.
- Production Engineering: Built real-time image processing pipelines, handling historical video streams with RabbitMQ and automated assembly.
- LLM Integration: Integrated multimodal models (Gemini, Pixtral) for semantic image analysis, robust JSON extraction, and contextual alerting.
- MLOps & Systems: Architected event-driven pipelines, containerized CPU/GPU environments (Docker), and established full observability with Grafana and GCP.
- Biometric Systems: Engineered high-accuracy Person Recognition systems using advanced Face Embeddings.
- Liveness & Security: Implemented Gaze Tracking and liveness detection algorithms to prevent spoofing.
- Documentoscopy: Specialized in Documentoscopy for fraud detection. Built robust pipelines for Brazilian Document Detection (RG, CNH) and automated data extraction via OCR.
- Orchestration: Managed MLOps infrastructure and deployed models for Classification, Object Detection, and OCR.
- Development: Built automated checking tools and RESTful APIs using TensorFlow, PyTorch, and AWS/GCP.
- Data & Cloud: Handled dataset annotation, cloud computing resources, and implemented logging policies with Elastic/Kibana.
