Machine Learning Scientist

About Us

Ultra is accelerating progress toward boundless automation through a grounded and applied approach to general-purpose robots.

Unlike traditional industrial automation that’s rigid and complex, our robots are zero integration—able to be installed in hours, not weeks—and are highly flexible, capable of quickly learning new tasks and delivering immediate ROI.

Founded by a team of three-time entrepreneurs with a decade of collaboration, Ultra moves fast and decisively. We already have robots in the field generating revenue and data, with plans to rapidly scale deployments this year.

We’re looking for an Machine Learning Scientist to join our NYC-based team (we are an in-person company), and help lead the development of our neural network robot controls, maximally leveraging our fleet of production robots and access to compute. We are an early stage company moving very fast in a rapidly growing space, and welcome people from any background as long as you’re excited to join our mission, drive immediate impact, and create a future where automation is accessible to all.

Who You Are

• Experienced with model training, deployment, and maintenance in a production environment

• Strong skills in deep learning and related tooling (e.g. Pytorch, Tensorflow, Ray, etc.)

• You know how to deploy high-quality production code as a part of a software team and care deeply about good engineering

• You thrive in a high-trust, high-autonomy environment. You don’t need to be micromanaged on what the top priorities are at any given moment.

• You’re hungry for impact and personal growth, and like to have fun in the pursuit

• You are a generalist who enjoys getting you hands dirty across the whole stack ML stack from R&D through production

• Deeply passionate about robotics and AI

What You’ll Do

• Improve our robots’ ability to quickly learn new tasks and to execute those tasks quickly and reliably

• Train VLA (vision-language-action) models for grasping and manipulation tasks on our robots using simulation and real-robot datasets

• Keep up with cutting edge research on robot learning (RL, imitation learning, etc.) and evaluate them to improve our stack

• Build pipelines to collect and curate data coming off our production robots

• Optimize our models for large scale training and low-latency edge inference

• Integrate trained models with our production robotics stack and work with deployment teams to scale models across varying warehouse configurations

Bonus Points