Senior Technical Product Manager - Serverless AI

Навыки

• API (интеграции)
• CUDA / ONNX / TensorRT
• Анализ данных
• Распределённые системы
• Docker
• HR-процессы
• Kubernetes

О компании и продукте

• Nebius 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.
• Built by engineers, for engineers. From large-scale GPU orchestration to inference optimization, we own the hard problems across compute, storage, networking and applied AI.
• Listed on Nasdaq (NBIS) and headquartered in Amsterdam, we have a global footprint with R&D 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 R&D.
• Nebius Serverless AI is our consumption-based compute platform for running AI workloads — training jobs, inference endpoints, and interactive development environments — without managing infrastructure. Users submit containerized workloads via CLI or UI, access GPU compute with pay-per-second billing, and the platform handles provisioning, lifecycle, and cleanup. We launched GA in Q1 2026 and are now scaling toward 1,000+ users while building the next generation of capabilities: autoscaling, multi-node distributed workloads, and developer-first tooling.

Требования

• Non-negotiables — you must have hands-on experience with
• You have built, shipped, and iterated on infrastructure or platform products used by developers or ML engineers
• Not consumer apps
• Not dashboards
• Infrastructure
• You understand containers at a practical level — Docker, image registries, container runtimes, resource limits, networking
• You've debugged why a container won't start, why GPU isn't visible inside it, or why a mount isn't working
• You have working knowledge of GPU computing for AI/ML: what GPU types exist and when to use them, how training and inference workloads differ in resource requirements, what vLLM / TensorRT-LLM / Triton are and why they matter
• You can read a CLI reference and know if it's well-designed. You've shaped developer-facing APIs, CLIs, or SDKs
• You have run real customer discovery — not surveys, but technical conversations with engineers where you learned something that changed your product direction
• You have 3+ years of product management experience in cloud infrastructure, AI/ML platforms, or developer tools
• Technical skills we will test in the interview
• Ability to whiteboard a workload lifecycle (submit → schedule → provision → execute → cleanup) and identify failure modes at each step
• Understanding of autoscaling trade-offs: scale-to-zero vs. warm pools, scaling metrics (queue depth, latency, utilization), cold start implications
• Familiarity with inference serving concepts: batching, model loading, quantization, KV-cache management, multi-model serving
• Understanding of distributed training concepts: data parallelism, model parallelism, communication overhead, checkpointing
• Ability to reason about pricing models: per-second vs. per-request vs. per-token, and how pricing interacts with product architecture
• It will be an added bonus if you have
• Experience at a serverless or GPU cloud company
• Hands-on ML engineering background — you've trained models, deployed inference endpoints, or built ML pipelines yourself
• Experience with Kubernetes for ML workloads (Kubeflow, KServe, Ray Serve) and understanding of why many ML teams want to avoid it
• Prior experience building a product from early stage to scale in a fast-growing market
• Background in systems engineering, distributed systems, or site reliability engineering
• Who thrives in this role
• You are more comfortable in a terminal than in a slide deck

Паспорт вакансии

История публикации

Откуда что взялось

Отмечено то, что вывели мы. Без пометки — значение назвал работодатель.

Почему на этом месте в выдаче

Порядок выдачи объявлен контрактом: свежесть решает между днями, полнота и зарплата — внутри дня.