Product Manager, Retrieval-augmented Generation and Embeddings

Google Google · Big Tech · San José, CA +2

Product Manager for Retrieval-Augmented Generation (RAG) and Embeddings, focusing on scaling these technologies across Google's AI-powered experiences. The role involves bridging infrastructure capabilities like chunking, inference, and embeddings retrieval, and guiding the evolution of embeddings across storage and serving systems.

What you'd actually do

  1. Partner with research teams across DeepMind, Research, and Core to advance the next-generation RAG technologies at Google, and make them available across the company.
  2. Partner with clients to unlock new business opportunities by resolving critical bottlenecks in Google’s RAG technologies.
  3. Drive the horizontal product outlook and road map to ensure smooth and efficient journeys across multiple products in the AI Foundations portfolio.
  4. Conduct client and external research to surface top client pain points and emerging opportunities.
  5. Define and track success metrics for user value and developer velocity.

Skills

Required

  • Product Management
  • Technical Product Management
  • Generative AI
  • LLM Interfaces
  • Developer Platforms
  • APIs/SDKs
  • Back-end Infrastructure
  • Search Applications
  • AI Applications

Nice to have

  • Master's degree
  • Software Development
  • Engineering
  • Vector Search
  • User Journeys
  • Success Metrics
  • Analytical Skills
  • Cross-functional Leadership

What the JD emphasized

  • Experience integrating generative AI tools or LLM interfaces into workflows.
  • Experience building and scaling developer platforms, developer-facing APIs/SDKs, or general back-end infrastructure (e.g. for search or AI applications).
  • Experience with Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), and vector search.

Other signals

  • LLMs need access to fresh, specialized data
  • Retrieval-Augmented Generation (RAG) setups to optimize cost, quality, and latency
  • vector (embeddings) search as a critical technology
  • guide the evolution of our embeddings capabilities across a number of storage and serving systems