Senior Software-Architekt

For a customer in the space sector in Spain, I led a team of two engineers to develop a bare-metal bootloader for a PCB with an ATmega128 MCU and 2 MB FRAM. The bootloader implemented an auto-correction mechanism using Hamming code to mitigate radiation-induced bit errors. Development was done in Linux (avr-gcc, avrdude) and Windows (Atmel Studio). Unit Testing and HIL Techniques implemented for quality assurance.

Smart City solution for a Spanish public sector customer. As Lead Developer and Architect, I led a 7-person team (scrum) to build and deploy into production machine learning models for traffic prediction developed with Apache Spark MLlib. The platform was based on microservices in Go and Python orchestrated with Kubernetes, Docker and Helm, storing sensor data in MongoDB (replica set) and integrating Apache Spark (cluster mode), RabbitMQ, Jupyter Server, ArcGIS (geo-JSON) and a custom Django/Python front-end. Monitoring was done with Grafana and Prometheus. APIs were exposed via REST (OAS) and MQTT, secured with Kong API Gateway (API Keys, OAuth). Deployment was on-premises with VMWare VMs. Development used Azure DevOps and GCP. A Flutter App was delivered for iOS, Android, and web.