Summary

Nebius is hiring a senior ML engineer for its Token Factory platform, focused on high-performance LLM inference, fine-tuning, and low-precision workloads across tens of thousands of GPUs. The role involves GPU profiling, inference optimization, custom kernels, and modern deep learning infrastructure; the working arrangement is not specified.

Responsibilities

• Profile gpu workloads with nsight and pytorch profiler
• Optimize llm inference throughput and latency
• Implement speculative decoding architectures in inference engines
• Build low-precision training and inference pipelines
• Optimize dense and moe model components
• Contribute to open-source inference engines
• Develop distributed systems and high-load services