Audio Inference Engineer, Model Efficiency

The Audio Inference Engineer for Model Efficiency at Cohere is responsible for optimizing audio model serving efficiency by enhancing core metrics such as latency, throughput, and quality. This role involves identifying bottlenecks in existing systems and implementing innovative solutions for audio processing and streaming workloads. Collaboration with training and serving infrastructure teams is essential to ensure seamless integration of model development and real-time audio inference deployment. Candidates should have significant experience in high-performance audio or machine learning inference systems, with proficiency in C++ and Python, and a focus on achieving results.