Physical AI Engineer

Looking to make impact with AI in the real world?

And would others describe you as fast-learning, persistent and hands-on?

Want to apply Read all the information about this position below, then hit the apply button.

About Captic

Captic builds AI vision kits for robotics and automated inspection. Our technology helps machine builders across Europe automate manufacturing tasks that traditional vision systems can't solve.

We're hiring 3 Physical AI Engineers (Hiring across levels, from strong juniors to senior engineers who've shipped vision into production) to build systems that work not just in simulation, but reliably on factory floors 24/7.

What you'll do

As a Physical AI Engineer, you'll work at the intersection of AI, computer vision, robotics and real-world machines.

You'll:

• Build and test AI vision solutions for robotics and inspection.
• Train, evaluate and improve computer vision models.
• Integrate cameras, sensors, robots and AI software.
• Debug real-world hardware and software problems.
• Run experiments and work with manufacturing data.
• Support customers across Europe.

This isn't a laptop-only job.

You'll work hands-on with cameras, robots and machinery in our lab and at customer sites.

What we need

• Strong Python and a real foundation in machine learning — these are the hard requirements
• Comfortable with Docker, Linux, and Ubuntu-based edge devices (Jetson, Raspberry Pi), or confident you can be within a month
• Basic networking literacy — enough to debug why the camera stopped talking to the Jetson
• Ownership mindset — you’re excited by the freedom, authority, and responsibility to solve problems end to end.

Nice to have, genuinely not required:

• Sensor integration or robotics platforms (ROS, robot arms, PLCs)
• VLA models — NVIDIA Isaac GR00T or similar
• Anything you've learned the hard way about industrial hardware

Why Captic?

You'll work on AI that interacts with the physical world, alongside a team where your work has a visible impact.

Instead of optimizing another benchmark or building features several layers removed from the end user, you'll get to see your software recognize, inspect and manipulate real objects on real machines.

If AI vision, robotics and manufacturing sound like a fun combination, we'd like to meet you.

How to apply?

Apply and show us what you've built

Send us a CV and one project link — a repo, a video, a write-up, anything. Ideally something physical that broke and that you fixed. xlxgzvr

Two honest sentences about what went wrong are more interesting to us than a polished README.