Descripción de la oferta
Merista builds the physical AI workforce for food production. This is our founding team robotics hire, reporting to the CTO, owning the robot’s mind, and everything between intent and motion. Full-time and on-site in Barcelona.
How do we control portion weight while meeting cycle-time requirements?
How do we collect demonstrations and production data that lead to measurable improvements?
You’ll help determine where it earns its place, how to evaluate it, and how it should work with sensing, control, and mechanical design.
You’ll lead technical choices across manipulation, learning, and control, working with perception and mechatronics engineers to define the interfaces. Your responsibility continues into deployment, where we find out which assumptions survive.
Execution and feedback control: trajectory generation, contact-rich manipulation, force control, and portion-weight feedback. You’ll connect policy outputs to real-time control, manage inference latency and timing, and define how the robot responds when an action cannot execute as intended.
You’ll build automatic recovery where practical and give a remote operator the context and control needed to intervene.
Manipulation policies and learning: deciding what the policy outputs, how it is trained and evaluated, and where imitation learning, reinforcement learning, simulation, or conventional control earn their place with limited fleet data.
Data and evaluation: capturing useful demonstrations, failures, and recoveries, and building repeatable tests that show whether changes improve accuracy, speed, and reliability.
Hands-on experience building and running control software on real robot arms, including trajectory execution and feedback from sensors. Practical understanding of control-loop timing, latency, and fault handling. Practical understanding of robot learning or related ML, including data quality, evaluation, and deployment.
Strong Python and enough systems understanding to debug across sensors, models, control, and hardware.
Interest in working closely with mechanical and mechatronics engineers.
Most meals in the developed world pass through a food factory. Factories cannot keep those rooms staffed.
Merista builds the physical AI workforce for food production: operated robots that take over these tasks. One robot rolls onto a factory floor without rebuilding the line. We deploy it and run it, and the factory pays for the work it does.
We've already built and operated a multi-site robotic fleet for restaurants, dealt with production failures, field operations, remote recovery, hardware iteration, deployments and customers. Our long-term vision is the self-driving food factory. Why join at this stage
Your work will shape the data we collect, how we measure progress, and what we deploy.
Our longer-term thesis is that useful factory deployments can create a compounding learning advantage: observe real failures, capture the right data, improve the system, and carry those improvements into the next deployment. Expect hardware delays, repeated experiments, and production debugging alongside new development. Part of the job is standing on a factory floor at 6am, watching a robot fail, and working with the team to understand what needs to change.
