Machine Learning Engineer

The role\nToken Factory is a part of Nebius Cloud, one of the world's largest GPU clouds, running tens of thousands of GPUs. We are building a high-performance inference and fine-tuning platform designed to push foundation models to their hardware limits. Our mission is to maximize throughput, minimise latency, and optimise cost-per-token across tens of thousands of GPUs.\n\nSome directions we are currently working on, and which you can be a part of:\nInference Optimization: Identifying LLM inference bottlenecks to drive production speedups. Squeezing the maximum performance for a wide range of LLM architectures at scale (e.g., GPT-OSS, Kimi K2.5, DeepSeek V3.1/V3.2, GLM-5).\nInference engines support: Implement novel speculative decoding architectures, optimise components of various LLM designs (dense/MoE, autoregressive/parallel), and contribute to open-source inference engines.\nLow Precision Training & Inference: Design and productionise low-precision (FP8, NVFP4/MXFP4) training and inference pipelines with measurable gains in throughput and cost-efficiency.\n\nWe expect you to have:\nA profound understanding of theoretical foundations of machine learning and transformer architecture.\nExperience profiling GPU workloads using Nsight, PyTorch profiler, or similar tools\nUnderstanding of GPU memory hierarchy and compute/memory tradeoffs\nFamiliarity with important ideas in LLM space, such as MHA, RoPE, KV-cache, Flash Attention, and quantisation\nUnderstanding of performance aspects of large neural network training (sharding strategies, custom kernels, hardware features etc.)\nStrong software engineering skills (we mostly use Python)\nDeep experience with modern deep learning frameworks\nProficiency in contemporary software engineering approaches, including CI/CD, version control and unit testing\nStrong communication and leadership abilities\n\nNice to have:\nExperience working with open-source inference engines (vLLM, SGLang, TensorRT-LLM), including contributions\nExperience with kernel languages or DSLs such as Triton, Cute, CUTLASS, CUDA\nA track record of building and delivering products (not necessarily ML-related) in a dynamic startup-like environment.\nStrong engineering skills, including experience in developing large distributed systems or high-load web services.\nOpen-source projects that showcase your engineering prowess\nExcellent command of the English language, alongside superior writing, articulation, and communication skills.\n\nBenefits & Perks:\nCompetitive compensation\nCareer growth and learning opportunities\nFlexibility and ownership\nCollaborative and innovative culture\nOpportunity to work on impactful AI projects\nInternational environment and talented teams