Product Manager, Core Machine Learning

At Google, we put our users first. The world is always changing, so we need Product Managers who are continuously adapting and excited to work on products that affect millions of people every day. In this role, you will work cross-functionally to guide products from conception to launch by connecting the technical and business worlds. You can break down complex problems into steps that drive product development.

This role is an opportunity to work on one of Google's Machine Learning (ML) Open Source Software products (OSS), serving both large internal genAI production use cases and working with other Machine Learning, Software (SW), Hardware (HW), and foundation model training companies.

In this role, you will work internally and externally to drive our Open Accelerated Linear Algebra (XLA) strategy with a focus on major ML hardware partners, external OSS communities like Triton, and AI/ML Engineers across the JAX and PyTorch ecosystem who want to use OpenXLA.

Google Cloud accelerates every organization's ability to digitally transform its business and industry. We deliver enterprise-grade solutions that leverage Google's cutting-edge technology, and tools that help developers build more sustainably. Customers in more than 200 countries and territories turn to Google Cloud as their trusted partner to enable growth and solve their most critical business problems.

Responsibilities:

• Drive improvements to Google's OpenXLA core developer experience, including Stable High-Level Operations (HLO) and PJRT, release quality, test and release, and community management.
• Enhance OpenXLA's Hardware Engineer (HW) partner evaluation and adoption through sustained engagements with executive technical leaders.
• Drive the development of improved XLA tooling, including monitoring, profiling, and debugging, for both internal and external use cases.
• Analyze PyTorch's compilation stack, including its strengths and weaknesses, and position and advocate OpenXLA as an alternative. Identify and incorporate hard-to-copy high value differentiation into OpenXLA as a result of your analysis.