Member of Technical Staff - Robot Learning

About the Company

At Machenta, we’re building intelligence that lets any robot learn any new skill as quickly as a human can: instantly, from a single example. We believe that achieving human-level adaptability in the physical world is a key milestone in the future of physical AI, and we’re already seeing glimpses of that future today. As a team, we’re united by this shared vision, and we work closely together in person to bring it to life. On the back of our $40m seed round, we’re now moving into a new office and lab space in the heart of King’s Cross, London. But we’re still right at the beginning here, and we’re looking for the right people to join us all for what lies ahead next. Robotics and machine learning span a diverse technical stack, and wherever you sit within it, everyone at Machenta takes ownership of their part of the company. If that kind of ownership resonates with you, and you’re excited about building together with an ambitious, closely-knit team, then we’d love to speak with you.

About the Role

As a Member of Technical Staff working on robot learning, you’ll be working on the core models that let our robots pick up new skills from just a single demonstration. You’ll design and run experiments on representation learning for in-context imitation learning, explore how the quality, quantity and diversity of data shape what our models can do, and help scale training and evaluation so we can test ideas quickly and at scale. It’s a research-driven but highly practical role: you’ll take ideas from first experiments to policies running on real robots, working closely with everyone at Machenta. If you’re excited by the moment a robot does something new after seeing it only once, you’ll find this work very rewarding.

In This Role You Will:

• Push the frontier of instant robot learning, enabling robots to acquire new skills from minimal demonstrations.

• Design and run experiments on representation learning for in-context imitation learning.

• Lead investigations into how data quality, quantity and diversity affect model performance, and turn findings into data generation strategies.

• Optimise and scale model training and evaluation pipelines across large datasets and compute.

• Deploy and evaluate policies on real robots to drive model improvements.

What We Hope You’ll Bring:

• Hands-on experience training and evaluating learning-based policies for robotics.

• A deep understanding of modern deep learning, including transformers, sequence models and generative policy classes such as diffusion or flow-based models.

• Strong Python and PyTorch skills, and experience writing clean, reliable code.

• Experience running large-scale training, including distributed training and efficient data pipelines.

• A rigorous, experiment-driven mindset: careful baselines, ablations and clear conclusions.

• Clear communication of your ideas and findings to the team.

Bonus Points If You Have:

• Experience with in-context learning, meta-learning or few-shot adaptation.

• Experience deploying learned policies on real robot hardware.

• Experience building or curating large robotics datasets, or working with large vision-language-action models.

• Publications at leading ML or robotics venues.

• A PhD in a relevant field.

• Experience at an early-stage robotics or AI company.

Join Us

We’re setting out to change how robots learn, and we know it’s an ambitious goal. If you’re excited by the idea of robots picking up new skills the moment they’re shown, and want to help build that future from the ground up, we’d love to hear from you. This role is based in person at our office in London.