Senior Machine Learning Engineer, LLM Inference Optimization

h3About Nebius /h3 pNebius is leading a new era in cloud infrastructure for the global AI economy. We are building a full-stack AI cloud platform that supports developers and enterprises from data and model training through to production deployment, without the cost and complexity of building large in-house AI/ML infrastructure. /p h3About Nebius /h3 pNebius is leading a new era in cloud infrastructure for the global AI economy. We are building a full-stack AI cloud platform that supports developers and enterprises from data and model training through to production deployment, without the cost and complexity of building large in-house AI/ML infrastructure. /p pBuilt by engineers, for engineers. From large-scale GPU orchestration to inference optimization, we own the hard problems across compute, storage, networking and applied AI. /p pListed on Nasdaq (NBIS) and headquartered in Amsterdam, we have a global footprint with RD hubs across Europe, the UK, North America and Israel. Our team of 1,500+ includes hundreds of engineers with deep expertise across hardware, software and AI RD. /p h3The role /h3 pAs a Senior Machine Learning Engineer on the Applied AI team at Nebius Token Factory, you will own model and endpoint optimization from model artifacts through production deployment. Your work will span model internals, quantization and model compression, speculative decoding, KV-cache optimization, inference engines, serving architecture, and benchmarking to improve latency, throughput, memory efficiency, GPU utilization, and cost per token while preserving quality and reliability. /p pYou will work with kernel and platform engineers to investigate serving problems, compare configurations, and resolve performance and quality regressions. You will evaluate model
- and engine-level optimizations together with distributed inference designs, including request routing, scheduling, prefill/decode coordination, and multi-node GPU execution. Your improvements will be validated through reproducible benchmarks and safe production rollouts under real-world workloads. /p h3Your responsibilities /h3 ul liOwn optimization projects for specific model families, customer endpoints, or serving backends. /li liCompare inference engines and recommend serving configurations suited to individual workloads. /li liDiagnose model-quality and performance regressions during production rollouts. /li liImprove LLM and VLM endpoint latency, throughput, memory efficiency, GPU utilization, quality, and cost per token. /li liDeploy, configure, benchmark, and extend engines such as vLLM, SGLang, TensorRT-LLM, Triton Inference Server, and NVIDIA Dynamo. /li liDevelop production model-compression workflows covering quantization, quantization-aware training, distillation, low-bit serving, and accuracy recovery. /li liImplement or integrate speculative decoding, draft-model approaches, KV-cache optimization, prefix caching, chunked prefill, continuous batching, and disaggregated prefill/decode serving. For prefill–decode disaggregation (PDD), evaluate KV-cache transfer, worker placement, and capacity balancing to determine which workloads benefit from the architecture. /li liDesign and improve LLM request routers and scheduling policies that balance worker utilization, queueing, request characteristics, and KV-cache locality while meeting latency and reliability targets. /li liScale dense and mixture-of-experts inference across multiple GPU nodes. Select and tune tensor, pipeline, data, and expert parallelism—including wide expert parallelism (WideEP)—with attention to hardware topology, expert load balance, and communication overhead. /li liBuild reproducible benchmark harnesses measuring TTFT, TPOT, tokens per second per GPU, p95/p99 latency, GPU memory, reliability, and cost per token. Validate architecture choices under representative traffic and consistent GPU budgets. /li liCollaborate with GPU kernel and platform engineers to trace bottlenecks across model code, kernels, runtime, scheduler, gateway, and cluster layers. /li liProduce design documents, performance reports, rollout plans, and customer-facing technical explanations. /li /ul h3Must-haves /h3 ul liStrong engineering skills in Python and PyTorch. /li liHands‑on experience deploying or optimizing LLM, VLM, or high‑throughput transformer inference. /li liPractical knowledge of at least one modern inference stack, such as vLLM, SGLang, TensorRT‑LLM, Triton Inference Server, NVIDIA Dynamo, Ray Serve, KServe, or an equivalent internal system. /li liA strong understanding of transformer inference bottlenecks involving KV cache, attention, memory bandwidth, batching, parallelism, and long‑context serving. /li liAbility to quantify trade‑offs among latency, throughput, quality, utilization, and cost. /li liExperience designing or optimizing distributed inference system architecture, with hands‑on work in one or more areas such as request routing, distributed scheduling, PDD, multi‑node in