Full-time

Researcher, Robotics

San Francisco, California, United States · On-site

$200,000-$400,000

Robotics Researcher, World Models & Sim-to-Real

Training robot policies mostly in simulation is cheaper, faster and safer than collecting everything on hardware. The catch is that a lot of what works in sim falls apart on a real robot. This team is working on closing that gap.

They’re a well-funded, research-led robotics startup in San Francisco. Their approach combines world models that can predict how a scene will change, simulation accurate enough to trust for training and testing, and a tight feedback loop with physical robots.

The role sits across all three. You’d move between models, simulators and hardware, rather than owning one narrow piece.

What you’ll work on:

  • Testing robot foundation models and policies in simulation, measuring how well they generalise across manipulation, navigation and planning tasks

  • Building the tooling to run that testing at scale

  • Training world models on multimodal robot data, including vision, depth, joint state and actions

  • Collecting and curating real-world datasets, then fine-tuning on a mix of simulated and physical data

  • Getting learned policies running on manipulators and mobile robots

  • Debugging whatever breaks along the way, whether it’s hardware, sensing, control or the model itself

You’ll have worked hands-on with physical robots and shipped software that runs on them. You’ll also be comfortable in at least one robotics simulator, such as Isaac Sim, MuJoCo or Gazebo, and write strong Python. Experience with robot learning, foundation models or world models is a big plus, as is being able to track a problem down across hardware, software and ML without getting stuck.

The founders are researchers from top academic and industry labs, and the senior team includes some well-known names in AI research. It’s a small team, so you’d have real influence over the technical direction and see your work reach real robots quickly.

📍 On-site, San Francisco

If it sounds relevant, we’d be happy to talk it through.

All applicants will receive a response.

Salary

$200,000-$400,000

Location

San Francisco, California, United States · On-site

Contract

Full-time

Your consultant

Nathan Hilliar

Full-time

Researcher, Robotics

San Francisco, California, United States · On-site

$200,000-$400,000

Robotics Researcher, World Models & Sim-to-Real

Training robot policies mostly in simulation is cheaper, faster and safer than collecting everything on hardware. The catch is that a lot of what works in sim falls apart on a real robot. This team is working on closing that gap.

They’re a well-funded, research-led robotics startup in San Francisco. Their approach combines world models that can predict how a scene will change, simulation accurate enough to trust for training and testing, and a tight feedback loop with physical robots.

The role sits across all three. You’d move between models, simulators and hardware, rather than owning one narrow piece.

What you’ll work on:

  • Testing robot foundation models and policies in simulation, measuring how well they generalise across manipulation, navigation and planning tasks

  • Building the tooling to run that testing at scale

  • Training world models on multimodal robot data, including vision, depth, joint state and actions

  • Collecting and curating real-world datasets, then fine-tuning on a mix of simulated and physical data

  • Getting learned policies running on manipulators and mobile robots

  • Debugging whatever breaks along the way, whether it’s hardware, sensing, control or the model itself

You’ll have worked hands-on with physical robots and shipped software that runs on them. You’ll also be comfortable in at least one robotics simulator, such as Isaac Sim, MuJoCo or Gazebo, and write strong Python. Experience with robot learning, foundation models or world models is a big plus, as is being able to track a problem down across hardware, software and ML without getting stuck.

The founders are researchers from top academic and industry labs, and the senior team includes some well-known names in AI research. It’s a small team, so you’d have real influence over the technical direction and see your work reach real robots quickly.

📍 On-site, San Francisco

If it sounds relevant, we’d be happy to talk it through.

All applicants will receive a response.

Your consultant

Nathan Hilliar

Full-time

Researcher, Robotics

San Francisco, California, United States · On-site

$200,000-$400,000

Robotics Researcher, World Models & Sim-to-Real

Training robot policies mostly in simulation is cheaper, faster and safer than collecting everything on hardware. The catch is that a lot of what works in sim falls apart on a real robot. This team is working on closing that gap.

They’re a well-funded, research-led robotics startup in San Francisco. Their approach combines world models that can predict how a scene will change, simulation accurate enough to trust for training and testing, and a tight feedback loop with physical robots.

The role sits across all three. You’d move between models, simulators and hardware, rather than owning one narrow piece.

What you’ll work on:

  • Testing robot foundation models and policies in simulation, measuring how well they generalise across manipulation, navigation and planning tasks

  • Building the tooling to run that testing at scale

  • Training world models on multimodal robot data, including vision, depth, joint state and actions

  • Collecting and curating real-world datasets, then fine-tuning on a mix of simulated and physical data

  • Getting learned policies running on manipulators and mobile robots

  • Debugging whatever breaks along the way, whether it’s hardware, sensing, control or the model itself

You’ll have worked hands-on with physical robots and shipped software that runs on them. You’ll also be comfortable in at least one robotics simulator, such as Isaac Sim, MuJoCo or Gazebo, and write strong Python. Experience with robot learning, foundation models or world models is a big plus, as is being able to track a problem down across hardware, software and ML without getting stuck.

The founders are researchers from top academic and industry labs, and the senior team includes some well-known names in AI research. It’s a small team, so you’d have real influence over the technical direction and see your work reach real robots quickly.

📍 On-site, San Francisco

If it sounds relevant, we’d be happy to talk it through.

All applicants will receive a response.

Your consultant

Nathan Hilliar

techire AI website footer image