Description
The Token Company is hiring a Member of Technical Staff, Research to own the end-to-end model training stack for compressing raw LLM inputs. The role involves designing and training models from scratch on NVIDIA B200s, covering data, architecture, training, evaluation, and production shipping of compression models. Candidates should have experience training models from scratch, strong ML fundamentals, transformer expertise, and a focus on shipping models that reduce real inference costs. The position offers a competitive base salary, significant equity, housing and other San Francisco benefits, visa sponsorship, and resources to build the research team.
