Summary
Philip Mardoum is a software engineer and computational neuroscientist with a decade of experience building predictive ML models and research-grade engineering for neural data and neurotechnology. He combines deep expertise in time-series signal processing, unsupervised learning, and deep reinforcement learning—work that spans retinal encoding models that outperformed prior state-of-the-art to industry-grade RL agents prototyped at Electronic Arts. At the University of Washington he wrote large-scale data pipelines, automated analysis and visualization tools, and designed a relational metadata database to support high-throughput electrophysiology and prosthetic research. He has a strong track record of translating academic research into usable systems, winning competitive grant funding and teaching courses for both university and public audiences. Now based in Seattle and currently at the Allen Institute, he brings a rare blend of neuroscience rigor and production-minded software engineering to problems at the intersection of biology and AI. An understated strength is his ability to move fluidly between hands-on coding, hardware interfacing, and high-level model design.
10 years of coding experience
2 years of employment as a software developer
PhD, Neuroscience, PhD, Neuroscience at University of Washington
BA, Biological Sciences, Computational Neuroscience, BA, Biological Sciences, Computational Neuroscience at University of Chicago