AI Software Engineer, Imitation Learning (f/m/d)

ABOUT TACTILIA

At TACTILIA, we are building the industry-ready robotic hand that closes one of the biggest gaps in Physical AI. Spun out of SCHUNK, the global market leader in gripping technology, we combine a decade of robotic-hand expertise and real industrial access with the speed and ambition of a deep-tech startup.

This isn't a research project waiting for its first customer. We already have a product, customers are ready to use it, and we are launching now. Your work will directly shape the electronics, actuators, and embedded systems that make that possible.

You will join at the moment when the hard engineering questions become real product decisions: how do we make a highly capable robotic hand reliable, manufacturable, serviceable, and simple enough to deploy on a real shop floor?

THE ROLE:

As an AI Software Engineer focused on Imitation Learning, you will develop the learning systems that enable TACTILIA's robotic hands to acquire dexterous manipulation skills from demonstrations.

You will work across data, machine learning and robotics to turn human demonstrations and robot trajectories into robust manipulation policies that can run on real hardware. Your work will help build the foundation for scalable skill learning and bring Physical AI from demonstrations into reliable real-world behavior.

WHAT YOU WILL DO:

• Develop imitation learning methods for dexterous manipulation and robotic hands
• Build data pipelines for collecting, processing and curating demonstrations and robot trajectories
• Develop and train policies from human and robot demonstrations
• Work with multimodal data including robot states, actions, vision and tactile signals
• Evaluate, debug and improve learned policies on simulation and real robotic hardware
• Develop methods for data efficiency, generalisation and robust policy execution
• Build software infrastructure for training, evaluation and deployment of learned behaviours
• Work closely with robotics and controls engineers to integrate learned policies with the robot's control stack
• Analyse failure cases and use them to improve datasets, training methods and policies
• Contribute to the development of scalable approaches for learning new robotic skills

WHAT YOU BRING:

• Degree in Computer Science, Robotics, AI, Electrical Engineering or a related field
• Strong software engineering skills in Python and preferably C++
• Experience with imitation learning, behavioural cloning or related learning-based robotics methods
• Strong understanding of machine learning and deep learning
• Experience working with robotic systems and real-world sensor or trajectory data
• Experience with PyTorch or a similar deep learning framework
• Good understanding of robot kinematics, dynamics and control
• Strong problem-solving skills and a hands-on approach to testing ideas on real systems
• Ability to work across machine learning, software and physical robotics

BONUS:

• Experience with dexterous manipulation, grasping or robotic hands
• Experience with teleoperation or demonstration data collection
• Experience with reinforcement learning or offline RL
• Experience with vision-language-action or multimodal policies
• Experience deploying learned policies on real robotic hardware

The next chapter of robotics won't be built in a lab. It will be built by people who care about what happens when technology meets the real world.

At TACTILIA, you'll work on one of the hardest open problems in robotics: giving machines the dexterity, reliability and robustness to interact with the physical world at industrial scale.

The technology is taking shape, the team is being built, and the standards around robotic hands and their skills are still open. That means your decisions will have a lasting impact — not just on a product, but on what comes next for Physical AI.

Build the hand. Define the standard. Shape what comes next.

If you're excited by the challenge of building robotic hands that works reliably in the real world, we'd love to hear from you.

We show an excerpt and a link, and the full text when the source allows it. Brands and logos shown belong to their owners.

Source: Statistisches Bundesamt (Destatis), table; dl-de/by-2-0; StarIn's calculation

Posted Nov 01, 2026

Posted Nov 01, 2026