Research Software Engineer, Multimodal AI

Google Google · Big Tech · San Jose, CA +1

Research Software Engineer focused on developing advanced AI for XR devices, specifically working with LLMs and multimodal agents. The role involves algorithm development, production-quality coding, comprehensive evaluation planning, and shipping the latest modeling innovations in areas like orchestration and multimodality.

What you'd actually do

  1. Develop algorithms/models to enhance AI agents for XR devices using techniques like prompting, few-shot learning, post-training techniques to improve model performance and real-world XR scenarios.
  2. Write production-quality C++/Python code and tests.
  3. Create a comprehensive evaluation plan, from dataset development to Key Performance Indicator (KPI) definitions and measurements.
  4. Identify, implement and ship the latest modeling innovations including, orchestration, multimodality, tool integrations, memory, hybrid agent architectures and personalization.
  5. Demonstrate concepts through rapid prototyping and iterative development, using team testing in close partnership with the XR product teams.

Skills

Required

  • software development
  • C++
  • Python
  • Generative AI
  • Machine Learning
  • deep learning
  • perception
  • computer vision
  • software design
  • software architecture
  • testing
  • maintaining
  • launching software products

Nice to have

  • JAX
  • TensorFlow
  • PyTorch
  • multimodal learning
  • large language models
  • AI agents
  • prompt engineering
  • few-shot learning
  • post-training techniques
  • evaluations
  • large-scale model training
  • deployment
  • technical leadership
  • Master's degree
  • PhD

What the JD emphasized

  • production-quality C++/Python code
  • comprehensive evaluation plan
  • latest modeling innovations
  • rapid prototyping
  • iterative development

Other signals

  • LLMs
  • agents
  • multimodal
  • XR devices