Description
Liquid AI, spun out of MIT, is building efficient AI systems and Foundation Models (LFMs) that operate on-device and at the edge. They are looking for someone with experience writing high-performance GPU kernels for training or inference, understanding low-level profiling tools, and integrating GPU kernels into frameworks like PyTorch. The role involves writing high-performance GPU kernels for inference workloads, optimizing architectures, implementing new techniques, and monitoring performance. San Francisco and Boston are preferred locations, but other locations are open.
