AI/ML Engineering Intern – Behavioral Safety (m/f/d)

Synapticon GmbH; Germany (OH); On-site; Full time

Job description

WHAT WE ARE LOOKING FOR

We're looking for a motivated AI/ML Engineering Intern (ideally 6 months, mandatory internship) to join our Safety AI team in Schönaich (close to Stuttgart), with a flexible start date. If you're excited about applying machine learning to genuinely open research questions in humanoid robotics — from semantic intent recognition to AI-based fall strategies — this is a rare opportunity to do research with a direct path to product.

HERE'S WHERE YOU MAKE AN IMPACT

• Prototype multimodal foundation model integration for robot safety context — exploring how large-scale models can give robots semantic awareness of their environment and the people in it

• Develop and test AI-based fall strategy controllers, working on one of the more genuinely open research questions in humanoid deployment

• Implement ML model inference in the robot control loop — bridging the gap between research models and the timing constraints of real-time robot software

• Benchmark behavioral safety classifiers for semantic intent recognition in simulation, building the evidence base for what these models can and can't reliably distinguish in deployment

• Build evaluation frameworks for AI safety behaviour aligned with ISO/IEC TR 5469 — contributing to a principled, documented approach to assessing AI in safety-relevant contexts

WHO YOU ARE

• Enrolled MSc student in Machine Learning, Computer Science, or Robotics — with a research mindset and the discipline to support it with rigorous, reproducible evaluation

• Strong PyTorch skills; hands-on RL experience using Stable Baselines3, Isaac Lab, or similar frameworks — you've trained policies and analysed why they fail, not just run examples

• You've integrated ML inference pipelines with ROS2 — or are confident you can make it work cleanly in a real-time control context

• Experience with foundation model fine-tuning is a strong plus — this work involves adapting large models for a specific purpose, not just deploying them off the shelf

• You're interested in safety-critical AI — not as a compliance checkbox, but as a genuinely hard technical and methodological problem

• Strong mathematical foundation across linear algebra, optimisation, and probability — you're comfortable with the theory behind what you implement

WHAT'S IN IT FOR YOU

• Flexible working hours when life takes unexpected turns

• Company support for gym memberships through Wellpass

• Birthday vouchers

• Drinks and coffee (free of charge, of course)

• Snacks and fruit for all employees, as well as breakfast every Monday

• Regular global team-building events such as boot camps, skiing, and summer BBQs

• An international team with long-term prospects

• Highly competent colleagues who are experts in their field and happy to support you in your learning and growth

• Modern office space

• Our company's WG (upon availability)