The AI Teams at Agile Robots are looking for a Machine Learning Engineer (m/f/d) VLA Training & Evaluation , who will build the frameworks for fine-tuning and benchmarking robot foundation models on collected robot data. The frameworks are meant to work across different foundation models and for teams with modest data and compute.
Responsibilities
- Training Workflows: Build reusable, model-agnostic training and evaluation workflows, with experiment tracking, model versioning, and checkpoint management, so that different foundation models can be swapped in without rebuilding the pipeline.
- Fine-Tuning: Implement fine-tuning of pretrained vision-language-action models on collected robot data, for example with parameter-efficient methods such as LoRA or adapters, or with distillation.
- Policy Training: Build training loops for action-chunking and transformer-based action models, diffusion policies, and VLM/VLA heads.
- Benchmarks: Build benchmarks with generalization splits, success metrics, and failure-analysis templates.
- Transfer Studies: Work with simulation engineers on studies, including metrics and ablations, that test whether improvements in simulation realism improve performance on the real robot.
- Research Baselines: Turn published methods into working baselines and templates that others can extend.
Essential Skills
- Academic Background: Degree in Computer Science, Machine Learning, Robotics, or a related field (Master’s or PhD typical), or an equivalent combination of formal training and demonstrated experience training deep learning models.
- Training Internals: Hands-on experience writing and debugging training loops in PyTorch or JAX, including data loading, loss design, and optimizer behavior.
- Fine-Tuning: Experience fine-tuning pretrained models and adapting them to new tasks and data.
- Robot Learning: Working understanding of how imitation-learned policies, such as action-chunking or diffusion policies, turn observations into robot actions.
- Experiment Management: Experience designing evaluations and ablations, and running experiments with tracking, versioning, and checkpoint management so that others can reproduce them.
Beneficial Skills
- VLA Models: Experience fine-tuning vision-language-action or vision-language models for robot control.
- Real-Robot Evaluation: Experience evaluating learned policies on physical robots.
- Sim-to-Real: Familiarity with measuring how well models trained in simulation transfer to real hardware.
What we offer
- A dynamic high-tech company, combined with financial soundness and world-class investors.
- Join an interdisciplinary, international team with 60+ different nationalities in a collaborative work environment.
- Lots of development opportunities as we continue to grow.
- Challenging tasks and impactful projects alongside experts that enable professional and personal growth.
- Corporate Benefits Program covering health, mobility, and learning for 100€ net per month.
- Modern office facilities with a rooftop terrace overlooking Munich, free drinks & fruits, and regular company events contribute to a good working environment
About us
Agile Robots SE is an international high-tech company based in Munich, Germany with a production site in Kaufbeuren and more than 3500 employees worldwide. Our mission is to bridge the gap between artificial intelligence and robotics by developing systems that combine state-of-the-art force-moment-sensing and world-leading image-processing technology. This unique combination of technologies allows us to provide user-friendly and affordable robotic solutions that enable intelligent precision assembly.
This is made possible by our employees, who bring out the best in each and every day with creativity and enthusiasm. Become part of this team and shape the future of robotics with us!
We are proud of our diversity and welcome your application regardless of gender and sexual identity, nationality, ethnicity, religion, age, or disability.
Find more English Speaking Jobs in Germany on Arbeitnow
Mostramos trecho e link por vaga, e o texto integral quando a fonte permite. Marcas e logotipos citados pertencem aos seus titulares.
