Applied Researcher

Aurora Innovation Aurora Innovation · Robotics · PITHQ · Software Autonomy Sensing

Research role focused on Vision-Language-Action (VLA) models, exploring and rebuilding architectures to address failure points in current state-of-the-art models for robotics applications. Requires PhD and expertise in VLM/VLA, with a focus on practical implementation and architectural creativity.

What you'd actually do

  1. Architectural Exploration: Deep dive into VLM/VLA architectures to determine their viability for our specific problem sets.
  2. Technical Mentorship: Act as the subject matter expert and guide for an exceptionally capable engineering team eager to move into this space.
  3. Practical Implementation: Bridge the gap between high-level academic theory and practical, scalable applications.
  4. Innovation: Redesign and optimize model architectures to leverage high-performance computing environments like Aurora.

Skills

Required

  • Python
  • C++

Nice to have

  • Vision-Language Models (VLM)
  • Vision-Language-Action (VLA) models
  • Robotics

What the JD emphasized

  • PhD Graduate in AI, Computer Science, or Robotics (Top-tier research lab background)
  • Deep expertise in Vision-Language Models (VLM) and Vision-Language-Action (VLA) models
  • Hands-on experience with 'jagged edges'—knowing where current architectures fail and how to troubleshoot them
  • Practical researcher: Ability to translate complex papers into production-ready code
  • Architectural Creativity: Ability to rethink and redo entire model architectures rather than just fine-tuning

Other signals

  • VLM/VLA architectures
  • rethink and redo entire model architectures
  • translate complex papers into production-ready code