Member of Technical Staff, Model Efficiency

A Member of Technical Staff for Model Efficiency at Cohere focuses on enhancing the performance of large language models (LLMs) by implementing optimizations that improve inference speed, latency, and throughput. The role involves deep technical work across the inference stack, diagnosing bottlenecks, and collaborating with modeling and systems teams to deploy performance improvements. Candidates should have strong programming skills in C++ or Python, experience with LLM inference ecosystems, and a background in performance optimization, particularly with GPUs and distributed systems.