Senior ML Engineer

The Token Factory team at Nebius Cloud is building a high-performance inference and fine-tuning platform to push foundation models to hardware limits. We are looking for a Senior ML Engineer with deep expertise in transformer architectures, GPU profiling, and large-scale model training and inference to help optimize throughput and latency while controlling costs. You will work on inference bottlenecks, speculative decoding architectures, low-precision training/inference pipelines (FP8, NVFP4/MXFP4), and contribute to open-source inference engines. Requirements include strong Python software engineering, experience with modern deep learning frameworks, knowledge of MHA, RoPE, KV-cache, Flash Attention, and familiarity with sharding strategies and kernel development. Nice-to-have items include experience with vLLM, Triton, CUTLASS, CUDA and building products in startup environments. The role offers competitive compensation, growth opportunities, and an international, collaborative culture. Remote work is available within Europe, with offices in Amsterdam, Berlin, London, and Prague. To apply, follow the Nebius careers portal instructions.