Senior Applied Scientist — Efficient LLM Inference & Optimization

Nebius; Palo Alto (CA); On-site; USD 195,200 - 262,200; Full time

Benefits offered by this job

Health insurance

401(k) plan

Parental leave

Remote work stipend

Life & disability insurance

Job summary

Nebius is seeking a Senior Applied Scientist to turn frontier inference bottlenecks into production-ready solutions. You will design rigorous experiments, write high-quality code in Python and PyTorch, and collaborate with ML engineers to ship research into production.

You will own well-scoped research projects, publish credible work, mentor engineers, and define evaluation methods for latency, throughput, and cost per token across LLM/VLM inference workloads.

Qualifications

• PhD in computer science, ML systems, or closely related field.
• Strong publication record or equivalent artifacts in ML/efficient inference/quantization.
• Strong coding ability in Python and PyTorch; move from idea to experiment quickly.
• Deep understanding of LLMs, VLMs, transformer inference, decoding, quantization, and production-serving tradeoffs.
• Strong experimental design skills, including ablations, baselines, metrics, and failure analysis.
• Excellent written and verbal communication.

Responsibilities

• Own focused research projects from hypothesis through experiment to production handoff.
• Prepare internal reports, technical blogs, or papers when externally credible.
• Partner with MLEs to ensure research prototypes become usable production components.
• Define and execute research programs in efficient LLM/VLM inference with measurable production impact.
• Invent, evaluate, and productionize methods for quantization, QAT, distillation, speculative decoding, KV-cache reuse and compression.
• Build high-quality prototypes in PyTorch, Triton, CUDA-adjacent tooling, or inference-serving frameworks; productionize them with engineers.
• Design rigorous evaluation methodology covering quality, latency, throughput, numerical stability, memory footprint, tail latency, and cost per token.
• Publish papers, technical reports, blogs, and open-source artifacts to build external credibility for Nebius Token Factory.
• Collaborate with MLE, GPU kernel, backend infra, product, and customer teams to select high-leverage research bets.
• Mentor engineers and scientists on experimental design and model/system tradeoffs.

Skills

Python programming

PyTorch

Research experience

Experiment design

Communication skills

Education

PhD in CS / ML / related field

Tools

PyTorch

Triton

CUDA

TensorRT-LLM

Job description

Nebius is seeking a Senior Applied Scientist to turn frontier inference bottlenecks into production-ready solutions. You will design rigorous experiments, write high-quality code in Python and PyTorch, and collaborate with ML engineers to ship research into production.

You will own well-scoped research projects, publish credible work, mentor engineers, and define evaluation methods for latency, throughput, and cost per token across LLM/VLM inference workloads.