[ My path ]
I’m Manu. I studied telecommunications engineering, specialised in electronics, and my first job, in 2015, was on a test bench in Seville, validating silicon: new imaging chips, one at a time. The bench was slow, so I automated it in Python and gave the team back more than 15 hours a week. Automating my way out of the manual part has been a habit ever since.
In Grenoble I moved into the camera itself: digital logic in VHDL and firmware in C for vision cameras that had to answer in under 10 ms. Then, in 2018, Shenzhen, as the engineer between our European teams and our customers in China. That is where I learned the part nobody teaches: a chip that works on your bench is not a product until it works in the customer’s system. The industry calls that job field application engineering, and it is still part of the job I do.
From 2019, working between Spain and the US, that job had a bottleneck: every new camera had to be integrated before a customer could even try it. So I built a platform on Kubernetes and AWS that turned out an SDK and a proof of concept per camera. Onboarding a customer went from 120 days to 20–30, and the infrastructure bill fell by 80%.
Today the test bench is a cluster I run myself, on hardware I own. You can see it running: what it serves, how it deploys and what it measures. My AI agents work on it, under written rules and the checks that enforce them. The rest of what I build is in projects, and what I learn along the way goes into the notes below.
I never switched worlds. I just kept moving up a layer: from silicon to firmware, from firmware to the customer’s system, and from there to the cloud. Every layer taught me the same thing: the work is done when it runs in someone else’s hands.
Outside work I’m part of Cloud Native Sevilla and NaN, a community of people building with AI, and I give the odd talk when I can. And I row.