At Atlas Codes, I develop agentic programming workflows that generate and test native iOS and macOS apps, as well as cross-platform desktop applications. The workflows use context engineering and UML specifications, with unit and E2E tests.
I also design and orchestrate autonomous OpenClaw agent groups for automation and data analysis, running in containers with actions subject to prior approval. My work includes implementing embedded language, vision, and vision-language models for on-device inference.
At BISECT LDA, I developed a Transformer-based deep learning model to predict video Quality of Experience, training and evaluating it with PyTorch. I also created a real-time inference pipeline in Rust for low-latency streaming environments.
My earlier work included configuring Linux-based operating systems and networks for an IoT cyber-physical systems Raspberry Pi cluster at FEUP, and prototyping sensor-triggered advertising kiosks and real-time interactive media pipelines at RSB. At Stone, Eventials, OSX, OGX, I developed Python dashboards, predictive analyses, and automated financial reports and KPI monitoring.
