Robotics Applied Scientist, Spatio-Temporal Scene Understanding & Memory

FieldAI • Irvine, CA

3d • Onsite

About The Position

FieldAI’s Irvine team develops risk-aware, reliable, field-ready AI systems for real robots, sensors, and field deployments. The Robotics Applied Scientist will research and build persistent spatio-temporal mission memory so robots can track what they have seen, where and when they saw it, and how it changes across missions and sites. This includes tracking assets that move, disappear, or degrade rather than repeatedly rediscovering them.

The role will build world models that allow robots to reason and plan over stored memory, and support recovery of physical parameters through interaction for physics-aware digital twins. The work combines learned and physically grounded models, using geometry and dynamics as ground truth, to bridge the perception-reasoning gap for long-term autonomy. Research is shaped by deployment needs, and methods are expected to run on real robots.

The position is based with FieldAI’s Irvine team, where systems are developed and tested on real hardware and sensors before deployment in unstructured, previously unknown environments. The role works across AI, software, robotics engineering, product, field deployment, and technical communication. FieldAI values diverse perspectives and describes its workplace as inclusive.

Requirements

• PhD in robotics, computer science, electrical engineering, aerospace engineering, mechanical engineering, or a related field.
• Publications at top venues such as RSS, CoRL, ICRA, IROS, T-FR, T-RO, CVPR, NeurIPS, or similar, on scene understanding, semantic SLAM, lifelong or spatio-temporal mapping, world models, or embodied reasoning.
• Hands-on experience with vision-language or foundation models applied to spatial reasoning, memory, or long-horizon embodied question answering.
• Depth in 3D representations, such as scene graphs, neural or Gaussian-splat maps, or object-centric maps, and the estimation underlying them.
• Strong Python skills.
• Ability to run methods on real robots using C++ and ROS 2.
• Ability to think clearly about the purpose of a representation and how to evaluate it.
• Comfort working in a fast-paced environment where research questions are driven by upcoming deployment needs.

Nice To Haves

• Change detection or long-term mapping results from real deployments.
• Real2Sim, system identification through interaction, or digital-twin work.
• Experience with behavior trees or task planning driven by language models.
• Industrial or field robotics exposure.

Responsibilities

• Investigate appropriate representations for persistent spatio-temporal robot mission memory across missions and sites.
• Build world models that enable robots to reason and plan over mission memory.
• Develop methods to recover physical parameters through interaction for physics-aware digital twins.
• Combine learned and physically grounded models, using geometry and dynamics as ground truth.
• Bridge the perception-reasoning gap for long-term autonomy.
• Run methods on real robots and help develop systems tested on hardware and improved through field deployments.

What This Job Offers

• Job Type: Full-time
• Career Level: Senior
• Education Level: Ph.D. or professional degree