Audio Inference Engineer, Model Efficiency

Audio Inference Engineer, Model Efficiency at Cohere. This role focuses on optimizing audio inference serving efficiency by improving latency, throughput, and quality for real-time and streaming audio workloads. The engineer will dive deep into systems, identify bottlenecks, and develop creative solutions, collaborating with training and serving infrastructure teams.

Skills and focus areas

• machine learning systems
• audio processing
• streaming workloads
• real-time inference
• model serving
• latency optimization
• throughput optimization