Tasks
Turning today's research into tomorrow's applications, together. At ZEISS, we focus on user-centric innovation to transform ideas into cutting-edge solutions. The ZEISS Innovation Hub @ KIT brings together students, researchers, and industry professionals to advance AI, computer vision, and machine learning in application-oriented projects. We are looking for highly motivated students to develop AI-based solutions for real-world vision tasks, including agricultural solutions, from data collection, simulation-based data generation, and model development to deployment on target devices. This internship provides hands-on experience across the full ML pipeline, including vision foundation models, data preprocessing and curation, quantization-aware training, edge-device deployment, and performance evaluation.
•Develop an efficient and reproducible workflow for application-oriented machine learning solutions
•Be involved in sample preparation and data acquisition in the laboratory for full data ownership
•Implement quantitative metrics to assess and optimize data preprocessing quality and guide relabeling efforts
•Curate, structure, and version datasets for reproducible experiments
•Build data preprocessing pipelines for AI and machine learning workflows
•Generate and evaluate simulation-based or synthetic data
•Run hyperparameter and model studies in a traceable environment
•Automate embedded-device tests and optimize accuracy and speed
Requirements
•Enrolled in a Bachelor's or Master's program in computer science, mathematics, physics, AI, or related fields
•Strong programming skills in Python and PyTorch
•Solid understanding of software development principles
•Fundamental knowledge of data preprocessing steps for computer vision and vision foundation models
•Familiarity with AI/ML concepts (e.g., supervised learning, deep learning architectures) is a plus
•Experience with edge-device deployment and model quantization is a plus
•Creative, pragmatic, and self-motivated with strong analytical skills
•Capable of working independently as well as in a team-oriented environment
•Excellent communication skills in English or German
•Passion for innovation, enthusiasm for new technologies, and motivation to contribute to interdisciplinary projects
Please provide your complete application documents (CV, transcript of records, etc.). Your ZEISS Recruiting Team: Selina Safradin
Conditions
•A dynamic and interdisciplinary environment
•Hands-on experience with real-world image, sensor, and simulation-based data, foundation models, model quantization, and edge deployment
•Learn how to set up a full machine learning pipeline
•Close mentorship from experts in deep learning
•An agile work environment
