Nebius
Staff Machine Learning Engineer, LLM Inference Optimization
- London
- Hybrid
- Senior
- Permanent
Owns a platform-level optimization pillar such as quantization, speculative decoding, KV-cache optimization, or long-context serving, taking models from artifact to fast, cost-efficient endpoints. Builds benchmarking and rollout frameworks, tunes engines like vLLM and TensorRT-LLM, and sets technical direction for MLEs, kernel and infrastructure engineers.
vLLM · SGLang · TensorRT-LLM +7