Take ownership of ambitious research and engineering directions, working with unique real-world data to advance robot learning.
About Mimetic Machines
Mimetic Machines builds learning and deployment environments for physical AI. We capture skilled work in real workplaces and build the data, tools, and environments that help robots learn to do useful work.
The role
We’re hiring Members of Technical Staff to take ownership of ambitious research and engineering directions at Mimetic Machines. We’re collecting a growing dataset of skilled physical work from industries and environments that are difficult to access. Unique real-world data with the potential to open new directions in robot learning.
We think this data can support breakthroughs in video understanding, world models, and robot learning that move robotics into useful work. If you have a compelling research bet in that direction, we want to hear it.
You’ll take ownership of a substantial part of our research and development, working directly with the founders. You’ll have access to the data, room to shape what we collect next, and the responsibility to carry an idea from its first experiment to a working system.
What you’ll do
- Define and lead a technical direction, with clear hypotheses, baselines, and measures of success.
- Develop, adapt, train, and evaluate models on real-world physical-work data.
- Build the data processing, tooling, and training infrastructure your experiments need.
- Turn promising results into reproducible systems that others can use.
- Investigate failures across the data, models, and underlying assumptions.
- Work closely with collection and operations so your findings improve what and how we capture.
What we look for
- You’ve built substantial technical work yourself: trained models, research systems, perception pipelines, or deployed robotics software.
- You have depth in at least one relevant area, such as computer vision, robot learning, simulation, or ML systems.
- You’re strong in Python and comfortable with PyTorch or an equivalent framework.
- You can turn an ambiguous objective into a concrete plan and make iterative progress independently.
- You care about whether results hold up on new data and under real conditions.
- You take responsibility for the outcome, including changing your approach when the evidence calls for it.
Experience that could be useful
Egocentric video, multi-view geometry, SLAM, hand-object interaction, imitation learning, vision-language-action models, scene reconstruction, physics simulation, dataset curation, or distributed training.
You don’t need experience in every area. We value depth, judgment, and the ability to learn and move quick.
Working at Mimetic
This is a founding role with room to shape both the technical agenda and how we work. You’ll stay close to the people, environments, and recordings behind the research. Your results will influence what we collect, what we build, and which directions we pursue.
We want someone who enjoys that responsibility and wants to grow with it.
#J-18808-Ljbffr
