Mana Robotics · Available immediatly · Dresden

Build the Physical AI powering the next generation of home robots.

The next major technological wave is approaching: Robots for the home. Technology is moving beyond the digital world and starting to take action in our physical lives – finally taking on the everyday tasks that smartphones and apps have barely been able to simplify.

Mana Robotics develops autonomous robots for everyday life, with the ambition of becoming Europe's leading home robotics brand: innovative, beautifully designed, trustworthy, and relentlessly focused on real-world usefulness.

Our first product tackles one of the most time-consuming household tasks: preparing meals every day. Our ultra-compact, autonomous cooking robot "Mana One" prepares meals from fresh ingredients entirely on its own – the biggest leap forward in household convenience since the invention of the dishwasher.

After several years of development, research in Physical AI, and eight generations of hardware tested in real-world environments, we are now bringing this new product category to market.

At the heart of our technology is a fundamental challenge: Teaching robots to perform complex physical tasks reliably in the real world. From handling ingredients and manipulating kitchen tools to adapting to unfamiliar situations, we believe learning-based robotics will be key to making truly autonomous home robots a reality.

That's where you come in!

Your Role

As a Machine Learning Engineer – Robotics, you will develop, train, and deploy learning-based models that enable our robots to perform complex manipulation tasks autonomously.

You'll work at the intersection of Robot Learning, LLMs, and real-world robotic systems, turning cutting-edge research into capabilities that work reliably on our proprietary physical hardware.

Together with our robotics and engineering team, you'll take ownership of the full Robot Learning pipeline – from collecting and preparing training data to experimenting with model architectures, evaluating performance, and deploying policies on our robots.

An ideal role for someone who is passionate about cutting-edge AI, enjoys solving challenging technical problems, and wants to see their models do more than generate predictions – actually interact with the physical world.

What You'll Do

You will initially focus on developing and advancing the skills of our autonomous cooking robot, with the opportunity to expand your work to entirely new categories of home robots as we grow. Your responsibilities will include:

• Robot Learning & Model Development: Research, implement, train, and improve learning-based models for robotic manipulation, including approaches such as Imitation Learning, Behavior Cloning, Diffusion Policies, Vision-Language-Action (VLA) models, and whatever groundbreaking approaches emerge next.
• Data Collection & Training Pipelines: Expand and improve workflows for collecting robotic demonstrations, processing multimodal sensor data, and creating high-quality datasets for model training.
• Model Training & Experimentation: Design, run, and evaluate experiments with modern learning architectures, pre-trained models, and fine-tuning techniques to improve robotic capabilities.
• Real-World Robot Deployment: Deploy trained policies onto our robotic systems and bring models from the lab to real-world user environments.
• Evaluation & Performance Optimization: Develop systematic evaluation methods, analyze model failures, and continuously improve the robustness, generalization, and reliability of learned behaviors.