← All roles

Software Engineer, Multimodal AI, XR

Google
Apply on company site
Hybrid (San Jose, CA; New York, NY; San Francisco, CA)$147k - $210kFull-timePosted 1w ago

About the role

Software engineer focused on advancing multimodal and conversational AI for next-gen wearable XR devices, developing on-device and hybrid models, and collaborating with XR product teams and researchers.

Responsibilities

  • Lead the development and optimization of on-device and hybrid multimodal models for XR devices. Utilize prompting, few-shot learning, and post-training (e.g., quantization) to enhance performance under power and latency constraints.
  • Create comprehensive evaluation plans for hybrid systems, from dataset development to KPIs measuring model accuracy and on-device efficiency.
  • Identify, implement, and ship the latest modeling innovations, focusing on hybrid agent architectures, edge-cloud orchestration, multimodality, tool integrations, and personalization.
  • Prove out concepts for on-device AI features through rapid prototyping and iterative development, enabling team testing and close partnership with XR product teams.
  • Work closely with research scientists, engineers, and product teams to drive the technical roadmap; share findings through conference publications and contribute to product launches.

Qualifications

  • Bachelor’s degree or equivalent practical experience.
  • 2 years of experience with software development in one or more programming languages, or 1 year of experience with an advanced degree.
  • 1 year of experience with one or more of the following: speech/audio, reinforcement learning, ML infrastructure, or another ML field.
  • 1 year of experience with relevant ML frameworks such as JAX, TensorFlow, or PyTorch.

Skills

Required:Programming in one or more languagesMultimodal AIOn-device model development
Last updated Oct 23, 2026. Interested? Apply on the company site.

About Google

Google Search homepage

llmtext-to-speechvoice-ai-stack
Company profileWebsite

More open roles at Google