Summary

Research and optimize efficient LLM and VLM inference, developing prototypes and evaluating trade-offs in quality, latency, throughput, memory, and cost. The role focuses on production-relevant inference systems and model optimization; no working arrangement is specified.

Responsibilities

• Lead research projects in efficient llm inference
• Develop and evaluate inference optimization methods
• Build prototypes with pytorch and triton
• Research request routing and serving strategies
• Investigate multi-node inference and expert placement
• Evaluate serving architectures under consistent workloads
• Share findings through reports and publications