Software Engineer Iii, Agentic Ecosystem Platform

Google Google · Big Tech · Belo Horizonte, State of Minas Gerais, Brazil

Software Engineer to build and scale core components of Google's Agentic Ecosystem Platform, focusing on third-party integrations and supporting agentic operations across major Google products like AIM and Gemini. The role involves designing, developing, and maintaining robust infrastructure for agentic operations, translating technical requirements into reliable solutions, and enhancing system scalability.

What you'd actually do

  1. Contribute to the design and implementation of core infrastructure for Agentic Integrations (with a focus on MCP Servers).
  2. Build domain-specific features leveraging generalist models (AIM/Gemini). Collaborate directly with vertical teams to execute tool orchestration, writing the critical code needed to ensure reliable and consistent user experiences.
  3. Write, test, and deploy surface-agnostic code that operates seamlessly across AIM, Gemini, and Geo. Implement well-defined components of the MCP2Agency bridge, ensuring all deliverables strictly adhere to data governance and privacy policies (e.g., SIAN).
  4. Deliver high-quality code for priority partner integrations within a globally distributed team.
  5. Build and maintain robust features that extend the search experience framework while upholding high standards for code and search quality.

Skills

Required

  • software development in one or more programming languages
  • Speech/audio (e.g., technology duplicating and responding to the human voice)
  • reinforcement learning (e.g., sequential decision making)
  • ML infrastructure
  • model deployment
  • model evaluation
  • optimization
  • data processing
  • debugging

Nice to have

  • building developer-facing platforms
  • software development kits
  • third-party integration frameworks
  • model context protocol (MCP)
  • Large Language Models (LLM) orchestration
  • tool-use (function calling)
  • agentic workflows
  • evaluations
  • model performance optimization

What the JD emphasized

  • core infrastructure for Agentic Integrations
  • tool orchestration
  • surface-agnostic code
  • data governance and privacy policies
  • priority partner integrations
  • search experience framework

Other signals

  • building and scaling core components of our Agentic Ecosystem Platform
  • delivering complex third-party integrations
  • translating technical requirements into highly reliable, production-ready solutions
  • owning critical platform features from end-to-end
  • contribute to the design and implementation of core infrastructure for Agentic Integrations
  • tool orchestration
  • surface-agnostic code that operates seamlessly across AIM, Gemini, and Geo
  • build and maintain robust features that extend the search experience framework