Product Manager II, ML Fleet Training

Google•Sunnyvale, CA

12d•$163,000 - $236,000

About The Position

As the Product Manager for ML Training Fleet Efficiency, you will lead product strategy and cross-functional engineering initiatives to improve utilization, efficiency, and governance across Google’s global AI accelerator fleet. You will guide products from conception through launch, connecting technical and business priorities and breaking complex problems into actionable steps. You will also advise executive leadership on sustainably powering Google’s next-generation AI work. The role is within Google Cloud, which provides enterprise-grade solutions and tools to organizations around the world. Product Management works with engineering, design, marketing, and other teams to guide products through execution, including analyzing, positioning, packaging, promoting, and tailoring solutions for users.

Requirements

Bachelor’s degree or equivalent practical experience.
5 years of experience in product management or a related technical role.
2 years of experience taking technical products from conception to launch, such as from ideation to execution or end to end.
Experience utilizing AI productivity tools to streamline individual, business, or development workflows.

Nice To Haves

Master’s degree or PhD in Computer Science or Computer Engineering, or an MBA with a technical undergraduate degree.
5 years of experience managing machine learning infrastructure, AI accelerator platforms (TPUs/GPUs), or high-performance computing (HPC) clusters.
3 years of experience developing and deploying AI systems, managing machine learning infrastructure, or evaluating and optimizing AI models for task execution and reliability.
Deep familiarity with modern machine learning training and post-training frameworks such as JAX, PyTorch, and Ray; distributed tracing; and data ingestion architectures.
Demonstrated success establishing company-wide metric standards, capacity governance models, or developer observability platforms.

Responsibilities

Define and execute the multi-year product roadmap for Google’s ML Training Fleet Efficiency across TPUs, targeting high-leverage bottlenecks and unlocking hundreds of millions in compute capacity.
Standardize company-wide efficiency metrics, including Prod System Duty Cycle (PSDC) and PADC, and telemetry architectures to provide actionable visibility to executive and engineering stakeholders.
Partner with DeepMind and Core ML to optimize distributed training workloads, particularly Reinforcement Learning, data ingestion pipelines, and checkpointing efficiency.
Drive productization of AI diagnostic platforms, such as agentic root-cause analysis and trace telemetry, to automate detractor resolution at fleet scale.
Lead resource governance and quota frameworks across Search, Ads, Cloud, and GDM, building executive partnerships to maximize accelerator utilization without compromising research velocity.
Work cross-functionally to guide products from conception to launch, connecting technical and business priorities.