adrian valverde

I’m an Applied Computer Vision, Edge AI, and Autonomous Systems engineer focused on high-performance analytics and reasoning-based orchestration for safety-critical environments. I optimize real-time perception and inference to deliver 10x speedups and sub-30ms latency using the NVIDIA Jetson ecosystem.

In my current role, I architect and develop an end-to-end, model-agnostic autonomy stack that shifts from passive monitoring to a reasoning-based feedback loop for smart infrastructure hazard mitigation and autonomous action execution. I build reliability into agent workflows with LangGraph, Model Context Protocol (MCP), strict schema validation using Pydantic, and integration testing with Pytest to stabilize perception-to-action loops.

I also design the supporting infrastructure for real-world operation—deploying Redis Stack for operational memory, orchestrating multi-channel telemetry via n8n, and executing hardware commands through MQTT. Beyond core perception, I develop multi-sensor industrial safety systems, smart mobility (VRU detection) pipelines, and stereo vision/3D reconstruction solutions for automated volumetric analysis.