Internship in Machine Learning, Visual Language (Action) Models

Welcome to ZEISS – a company that combines innovation and responsibility! Our corporate functions are diverse and make a decisive contribution to the strategic orientation and sustainable success of ZEISS.

We are seeking passionate and talented students who want to make an impact by helping to develop next generation products and innovative machine learning algorithms at ZEISS.

Integrated in a fast-growing team of scientists and research engineers, you will work (for at least 3 months (ideally 6) , starting as early as possible, preferably already from August or September ) on helping us to develop core algorithms and taking academic ideas into working prototypes.

Your role With us, you have the opportunity to perfectly combine your studies with practical experience while actively contributing to exciting projects.

This allows you to gain valuable skills, expand your network, and grow both professionally and personally. • Work with the latest VLMs and / or VLAs • Read latest research on the topic, implement and use SOTA training techniques • Evaluate results qualitatively, come up with metrics to do quantitative evaluations • Plan and conduct experiments, benchmark and analyze results • Document your findings Your profile • Bachelor’s degree in Computer Science, Mathematics, Physics, or related fields with an outstanding study record, and currently enrolled in a master’s program • Solid understanding of Machine Learning and Computer Vision • Solid understanding of transformers, LLMs and VLMs • Programming background in Python and hands-on experience with Pytorch • Effective communication and presentation skills in English ​ Ways to stand out: • Industrial / academic experience with VLMs / VLAs, publications on the topic • Solid understanding of modern Reinforcement Learning • Solid understanding of Flow Matching Sounds exciting?

Then become part of #teamZEISS and help us shape the future! Please provide your complete application documents (CV, transcript of records) .