Senior Technical Product Manager - Serverless AI

# Senior Technical Product Manager – Serverless AI

## About the role

Nebius is seeking a Senior Technical Product Manager for the Serverless AI product team. You will share ownership of the product – but have individual full autonomy over your areas. This is not a role where you write and hand off requirements. You are the person who deeply understands container runtimes, GPU scheduling, cold start optimization and inference serving to make correct technical trade-offs – and also the person who talks to customers, shapes the CLI experience, sets pricing and drives adoption.

## Responsibilities

### 1. Product Ownership

• Co-ownership of the Serverless AI product roadmap (Jobs, Endpoints, DevPods)
• Write detailed, technically precise PRDs that engineering teams can execute
• Make build/buy/defer decisions on capabilities such as autoscaling, multi-node orchestration, HTTPS termination, secret injection and health checking

### 2. Technical Depth

• Understand the complete workload lifecycle: container image pull → VM provisioning → GPU attachment → workload execution → cleanup
• Evaluate technical trade-offs in container cold start optimization, GPU scheduling and storage mount performance
• Work directly with engineers on architecture decisions for distributed training support and endpoint autoscaling
• Stay informed about the fast-moving serverless GPU infrastructure space

### 3. Customer & Market

• Conduct customer discovery and feedback sessions with ML engineers and platform teams
• Analyze usage data, activation funnels and churn patterns
• Track market dynamics and emerging technologies
• Define and iterate pricing, packaging and tier strategy

### 4. Go-to-Market

• Own technical content strategy: quickstart guides, tutorials, reference architectures
• Partner with marketing on developer-focused campaigns
• Collaborate with Solution Architects and Sales

## Success in 12 months

• Clear product-market fit with improving activation and retention metrics
• Multi-node jobs and autoscaling endpoints implemented
• Cold start time reduced from 1-3 minutes to under 60 seconds
• Developer experience (CLI, docs, error messages) sets the standard
• At least 3 product decisions directly attributable to customer conversations