Computer Vision / AI Intern (m/f/d)

Focus: Real-world vision and tracking systems – from dataset work to production-ready models

Location: Berlin / Hybrid

Start: As soon as possible or by arrangement

We are a Berlin-based defense deep-tech startup developing mission-critical systems for defense against drones, missiles, and other airborne threats.

Our systems detect, track, and support the neutralization of airborne objects in complex real-world environments.

We operate at the intersection of computer vision, AI, sensing, and embedded software. As a fast-growing startup, we combine rapid execution with high technical standards. You will work close to the product, with real data, and on systems that must perform reliably under demanding conditions.

As a Computer Vision / AI Intern, you will contribute to the development of real-world vision and tracking systems — from dataset creation and curation to production-oriented models.

This is a deliberately hands-on role. Dataset work and labeling are not side tasks, but a central part of the position and critical to model performance.

• Contribute to the development and improvement of models for detection, classification, and tracking of airborne objects
• Build, maintain, and analyze datasets based on real-world image and multi-sensor data
• Perform data labeling, label review, and help define robust labeling standards as a core part of the work
• Run and document training experiments, tests, and model comparisons
• Investigate edge cases, domain shifts, and model failures on real operational data
• Support data preprocessing, augmentation, and experiment tracking
• Contribute to validation of models under realistic conditions and preparation for edge or embedded deployment

• Advanced studies in Computer Science, Data Science, Robotics, Electrical Engineering, Mathematics, or a related field
• Initial experience with machine learning or computer vision through coursework, projects, research roles, or personal work
• Strong Python skills; experience with PyTorch, TensorFlow, OpenCV, or similar tools is a plus
• Interest in practical, production-oriented development rather than purely theoretical research
• Attention to detail in dataset work and willingness to treat data labeling as a key driver of model quality
• Proactive mindset, strong learning ability, and motivation to work in a small, high-performance team

• A paid internship with real responsibility rather than purely supportive tasks
• Insight into a fast-growing defense deep-tech startup with strong technical depth
• Close collaboration with experienced engineers and the core team
• Work on real data, production-oriented models, and edge-adjacent systems
• High ownership, steep learning curve, and the opportunity to make visible impact
• Hybrid work model, flexible hours, and office in Berlin

• Hands-on and pragmatic: You enjoy working with real data and real systems
• Detail-oriented: You understand the importance of data quality for model performance
• Ownership-driven: You take responsibility and push your work forward
• Curious and driven: You are eager to learn and grow quickly
• Collaborative: You enjoy working across teams including vision, tracking, sensor fusion, embedded, and hardware
• Mission-aligned: You are motivated to contribute to impactful, real-world technology

• You will work on real vision systems rather than purely theoretical demos
• You gain exposure to the full pipeline: from dataset creation and labeling to model training and deployment
• You collaborate closely with founders and interdisciplinary teams across vision, tracking, sensor fusion, embedded systems, and hardware
• You take on responsibility early and grow rapidly in both technical depth and practical impact