Summary
Milo Sobral is a research scientist based in Montreal with 8 years of experience building and deploying deep learning systems at the intersection of neuroscience and edge computing. He led development of Portiloop, achieving sub-250ms closed-loop EEG stimulation on edge hardware, and created subject-adaptive algorithms that improved sleep-spindle detection in multi-lab studies. Milo combines strong engineering chops—from C/C++ and Unity tools for large-scale annotation to PyTorch pipelines for online and on-device inference—with academic rigor from McGill and Polytechnique Montréal. Currently at Mila working on neuro foundational models, he has a track record of turning experimental AI research into robust tools used by dozens of human-subject experiments. Notably, his work bridges low-latency embedded systems and adaptive deep learning, enabling real-world neuroscience applications.
8 years of coding experience
2 years of employment as a software developer
Exchange Semester, 5.75/7.0, Exchange Semester, 5.75/7.0 at The University of Queensland
Master's degree, Artificial Intelligence, Master's degree, Artificial Intelligence at Polytechnique Montréal
Bachelor of Science - BS, Computer Science, 3.72/4.0, Bachelor of Science - BS, Computer Science, 3.72/4.0 at McGill University
Spanish, Portuguese, English, French