Engineer on the R&D team at OMI, building full-stack software for systems that touch the physical world — electric and hybrid off-road vehicles, and the production lines that build them.

Right now I'm building a traceability system for a new production line: React on the front with unit and integration tests, Python and Flask on the back, and the service layer in between that talks to the equipment on the floor. That's the part I find interesting — a retry isn't free when the second one moves a piston, so the whole thing has to be idempotent and fail gracefully.

My first job was teaching kids to build video games in GameMaker Studio 2. These days I'm studying an M.Sc. in Applied Artificial Intelligence at Tec de Monterrey.

smams — Real-time Mexican Sign Language recognition. A CNN in TensorFlow/Keras over 27 classes, trained on a dataset we captured ourselves. Nine model iterations took it from 28% to 82% validation accuracy — mostly by tripling the dataset, not by making the model cleverer.

BayTec — An assistive medical robot on a LoCoBot WX250s: autonomous navigation, face recognition to verify the patient, vital signs over MQTT, and medication delivery.

Teleoperative Assistant Robot — A robotic arm driven by hand gestures from a Myo armband, over a socket link. First place at Expo Ingenierías 2022.

📍 Puebla, México · LinkedIn · anthony.ibrail@gmail.com