Cameron Craddock

I'm a technical lead at Meta building production machine learning systems for low-power wearable devices and neural interfaces. I lead cross-organizational work to measure, predict, and optimize power, latency, and memory for always-on neural network inference.

I've architected high-resolution power measurement and automated performance-testing infrastructure, integrating power, latency, and memory feedback into development and CI. My work reduced neural-network execution time from 25 ms to under 8 ms, active-inference power by 42%, and system power by 20%.

Previously, I led recommender-system and personalization work at Indeed, managing six engineers and establishing data-science capabilities for Career Guide. At Dell Medical School, I led federally supported AI and computational neuroimaging research, open-source software development, and large-scale biomedical data systems.

I also co-founded Brainhack and founded the Preprocessed Connectomes Project to advance open, reproducible neuroscience. Across industry and research, I build practical ML, embedded, and infrastructure systems that turn complex models into reliable products.