Software Engineer Iii, ML Pipeline Development, Xr

Google Google · Big Tech · Zürich, Switzerland

Software Engineer III, ML Pipeline Development, XR role focuses on productionizing advanced models and pipelines for on-device and server-based XR applications. Responsibilities include developing algorithms and frameworks for end-to-end pipeline development, optimizing system performance for hardware constraints, and ensuring critical power KPIs are met. Training ML models is not a requirement.

What you'd actually do

  1. Lead the productionization process by preparing models and pipelines for deployment.
  2. Develop algorithms and frameworks to support end-to-end pipeline development for on-device and server environments.
  3. Optimize system performance specifically for low-power and high-performance hardware constraints.
  4. Execute detailed implementations and conduct thorough power breakdown analysis.
  5. Ensure the achievement of critical power Key Performance Indicators (KPIs) necessary for the deployment of real-time spatial technologies.

Skills

Required

  • Python or C++ programming languages
  • computer vision (e.g., image classification and processing, object detection, visual search), video generation, or signal processing
  • ML infrastructure (e.g., model deployment, model evaluation, optimization, data processing, debugging)

Nice to have

  • building and optimizing production-level pipelines for 3D computer vision or machine learning models
  • on-device and server-side deployment frameworks
  • developing and deploying software for low-power, high-performance hardware
  • perform power breakdown analysis and optimize for strict hardware KPIs

What the JD emphasized

  • productionization
  • on-device and server-based platforms
  • optimization
  • system delivery
  • XR
  • low-power and high-performance hardware constraints
  • critical power Key Performance Indicators (KPIs)

Other signals

  • productionization
  • on-device and server-based platforms
  • optimization
  • system delivery
  • XR