Henry Ndubuaku is Co-Founder and CTO of Cactus Compute (YC S25), building a low-latency, open-source AI inference engine that runs large language models and vision models directly on smartphones, wearables, and edge devices.

Henry holds an MSc in Artificial Intelligence from Queen Mary University of London, started university at 15 studying EECS, and earned a Google GADS Scholarship via Andela before completing his graduate studies.

Henry turned down an AI/ML role at Nvidia to build Cactus, driven by a conviction that the future of AI is private, local, and on-device rather than cloud-dependent.

Prior to Cactus, Henry served as an AI Research Engineer at Deep Render working on real-time video models for phone GPU/NPU, and as an ML Software Engineer at Wisdm building distributed perception AI for Maxar Defence satellite imagery.

Cactus has attracted backing from Y Combinator, Oxford's Seed Fund, and FCVC, with research co-maintained by groups at UCLA, Yale, UPenn, Imperial College, and Stanford, and has published papers accepted at ICLR 2026.