Machine Learning Engineer

Primis; Columbus (OH); On-site; USD 120,000 - 180,000; Full time

Job summary

Primis in Columbus, OH is seeking a Machine Learning Engineer to build the learning systems behind autonomous industrial robots. This hands-on, deployment-first role moves behavioral cloning, reinforcement learning, diffusion and vision-language-action models from research to live robotic cells running in production.

You'll work with real-world vision and force/torque data, integrating manipulation and planning into production platforms and collaborating with hardware and software teams to push

Qualifications

• PhD or Master's in robotics, mechanical/electrical engineering, CS or related field.
• Publication track record at top venues (NeurIPS, ICML, etc.).
• Hands-on ML for robot manipulation essential.
• Experience deploying scalable autonomous robot systems.
• Proficiency in Python and/or C++, and ability to explain ideas clearly.

Responsibilities

• Develop and deploy learning-based methods for robot control and manipulation across robotic cells.
• Utilize multi-modal data (images, force/torque) to improve reliability and adaptability.
• Design/build vision-language-action models; manage data collection and fleet learning.
• Collaborate with hardware/software teams to integrate planning algorithms into production.

Skills

Hands-on ML for robot manipulation

Python/C++ programming

Deploying ML to real-world robotics

Communicating complex ideas

Experience with robotics simulation

Pose estimation / perception / bin-ptr

Education

PhD or Master's in Robotics / related field

Tools

Isaac Sim

MuJoCo

Job description

Machine Learning Engineer | Robot Learning

Build robots that learn from the real world

We're looking for a Machine Learning Engineer to build the learning systems behind autonomous industrial robots. This is a hands-on, deployment-first role. You'll take behavioral cloning, reinforcement learning, diffusion and vision-language-action models from research all the way to a fleet of robotic cells running in live manufacturing environments. Your models will learn from real-world vision, force and torque data, not just simulation.

What you'll be doing

You’ll develop and deploy learning-based methods for robot control and object manipulation across a fleet of industrial robotic cells. You’ll use multi-modal data, including images, force and torque, to make robots more capable, reliable and adaptive. You’ll design and build vision-language-action (VLA) models, and shape strategy across the whole solution lifecycle: data collection, fleet learning and continuous learning. You’ll work closely with hardware and software teams to integrate manipulation, control and planning algorithms into production robotic platforms, and help keep the team at the frontier of robot learning.

What we’re looking for

• PhD or Master's in Robotics, Mechanical Engineering, Electrical Engineering, Computer Science or a related field


• A track record of publishing at top venues (e.g., NeurIPS, ICLR, ICML, CoRL, RLC)


• Hands‑on experience applying machine learning to robot manipulation


• Experience building models from scratch and deploying them in real-world applications, not just research or sim


• A proven track record of deploying scalable solutions for autonomous robotic systems


• Experience with robotic simulation tools such as Isaac Sim or MuJoCo


• Experience with pose estimation, segmentation, active perception and affordance-based bin picking


• Strong programming skills in Python and/or C++


• Clear communication: you can explain complex technical ideas to a wide range of audiences

The bottom line

If your background is real robot learning, real hardware, and models that leave the lab and do real work, and you want your policies improving every day on a fleet of production robots, we’d love to talk.

We encourage candidates from underrepresented groups to apply even if you don’t check every box. Research shows women and underrepresented candidates often leave themselves out of roles they’re qualified for. If this sounds like your kind of problem, we want to hear from you.