Senior Machine Learning Engineer, LLM Inference Optimization

The role 

Nebius Token Factory is building fast, reliable, and cost-efficient inference services for frontier models. As a Senior Machine Learning Engineer on our Applied AI team, you will own model and endpoint optimization from model artifacts through production deployment. Your work will span model internals, inference engines, serving architecture, and benchmarking, with a focus on improving latency, throughput, memory efficiency, GPU utilization, and cost per token while maintaining model quality and reliability.

This is a hands-on role in which you will work on complex optimization projects, diagnose difficult serving problems, and deliver measurable improvements in production. Working closely with kernel and platform engineers, you will evaluate serving configurations, resolve performance and quality regressions, and optimize inference for real-world workloads, supported by reproducible benchmarks and safe production rollouts.

Your responsibilities: 

• Own optimization work for specific model families, customer endpoints, or serving backends.

• Run engine comparisons and recommend practical serving configurations for specific workloads.

• Debug model quality or performance regressions during production rollouts.

• Optimize LLM and VLM endpoints for latency, throughput, memory efficiency, GPU utilization, quality, and cost per token.

• Deploy, configure, benchmark, and extend inference engines such as vLLM, SGLang, TensorRT-LLM, Triton Inference Server, NVIDIA Dynamo, or similar systems.

• Build and productionize model-compression workflows, including quantization, quantization-aware training, distillation, low-bit serving, and accuracy recovery.

• Implement or integrate speculative decoding, draft-model approaches, KV-cache optimization, prefix caching, chunked prefill, continuous batching, and disaggregated prefill/decode serving.

• Build reproducible benchmark harnesses for TTFT, TPOT, tokens per second per GPU, p95/p99 latency, GPU memory, reliability, and cost per token.

• Partner with GPU kernel engineers and platform engineers to diagnose bottlenecks across model code, kernels, runtime, scheduler, gateway, and cluster layers.

• Write clear design docs, performance reports, rollout plans, and customer-facing technical explanations.

Must-haves: 

• Strong Python and PyTorch engineering skills.

• Hands-on experience deploying or optimizing LLM, VLM, or high-throughput transformer inference systems.

• Practical knowledge of at least one modern inference stack such as vLLM, SGLang, TensorRT-LLM, Triton Inference Server, NVIDIA Dynamo, Ray Serve, KServe, or equivalent internal systems.

• Strong understanding of transformer inference bottlenecks, including KV cache, attention, memory bandwidth, batching, parallelism, and long-context serving.

• Ability to reason quantitatively about latency, throughput, quality, utilization, and cost tradeoffs.

• Strong communication skills and ability to collaborate with research, kernel, infrastructure, product, and customer teams.

Nice-to-haves: 

• Experience with quantization-aware training, post-training quantization, FP8, INT8, INT4, NVFP4, MXFP4, AWQ, GPTQ, SmoothQuant, or related techniques.

• Experience with distillation, speculative decoding, EAGLE, Medusa, multi-token prediction, or other inference acceleration methods.

• Experience with agentic workloads, including tool calling, structured outputs, streaming APIs, high concurrency, and multi-step orchestration.

• CUDA or Triton familiarity, even if the role is not primarily a kernel-engineering role.

• Open-source contributions to vLLM, SGLang, TensorRT-LLM, FlashInfer, LMCache, PyTorch, Triton, Ray, KServe, or related projects.