AI Research Engineer at Cactus (S25) $120K - $180K • 0.10% - 1.00% AI inference engine for phones & AI-native devices San Francisco, CA, US / Remote (San Francisco, CA, US) Full-time US citizenship/visa not required Any (new grads ok) About Cactus Cactus is a cross-platform open-source framework for running AI models locally on smartphones. Cactus completed a highly oversubscribed fundraising round and now actively scaling the engineering team. About the role Skills: Torch/PyTorch, C++, Python, Machine Learning, Deep Learning Company Description Cactus is a cross-platform open-source framework for running AI models locally on smartphones. Cactus ships Kernels & AI engine for running AI locally on mobile devices; CUDA for phones. App developers can now deploy AI models locally in their mobile apps. Cactus completed a highly oversubscribed fundraising round and now actively scaling the engineering team. Role Description AI Research Engineers read papers and replicate models/techniques in Cactus. They also build and maintain Python tools for porting JAX/PyTorch models to Cactus graphs. They synergize strong mathematical foundations, intimate understanding of machine learning frameworks and experience with foundation models. AI research engineers also work with model companies to get them started with Cactus. This role requires ownership and high sense of agency. Qualifications Masters/PhD degree in CS/Maths/AI/Physics, or relevant experience Strong background in Python, PyTorch/JAX internals Experience with deep learning and SOTA LLM/VLM/TTS Experience with C++ and ml inference is preferred Ability to work collaboratively in a team environment A sense of urgency and interest in startups Benefits: $120,000 - $180,000 base salary Generous equity based on experience Health insurance, gym membership, etc. Work at a YC startup experience Cactus Repo: https://github.com/cactus-compute/cactus Technology Cactus ships Kernels & AI engine for running AI locally on mobile devices; CUDA for phones. App developers can now deploy AI models locally in their mobile apps.