At SWARM Biotactics , we’re creating biorobots — living, intelligent systems based on insects. Each insect carries a custom-designed backpack that enables wireless control, sensing, and secure communication.

Biorobot swarms combine the agility of biology with AI, sensors, and swarm coordination — unlocking capabilities where conventional machines fail.

We are the only company in the world doing this — and we’re pushing the edge of what’s possible in every domain: from biology to embedded hardware to swarm intelligence.

Effective and efficient environmental recognition is crucial for achieving our goals. By building a model of the environment, we enable focused action of the swarm and its individuals. We are looking for a motivated and driven working student (full- or part-time) to help us advance one of our environmental recognition pipelines. The objective is to enhance the pipeline’s capabilities. This can be achieved by improving existing capabilities. However, a key part of this work is to add new capabilities. In some cases, these capabilities already exist on other systems and simply need to be migrated. We strongly encourage adding your "own" new capabilities that don't yet exist in our stack.

This role involves work in the fields of Software Engineering, Machine/Deep Learning, and DevOps. You will work with multimodal data, SOTA Machine/Deep Learning inference frameworks, and SOTA edge deployment solutions. We do not require in-depth knowledge in all of these areas. A strong work ethic and the capacity to learn the necessary skills are essential. We're looking for someone with a strong sense of ownership for “their” product and a desire to keep improving it.

Things you could work on

• Improve the existing functionality of the existing environment recognition pipeline by employing other or more analysis methods, switching the inference framework, and enhancing usability
• Migrate existing capabilities into the environmental recognition pipeline by adding new sensor inputs, processing and analyzing them, and providing results to other systems
• Add new Machine/Deep Learning based functionality (e. g. SLAM) to the environmental recognition pipeline
• Reduce data transmission footprint by implementing modality specific compression, data encoding and improved analysis process design
• Possibly redesigning the system to fit new requirements like multi-agent inputs
• CI/CD with containerization and testing

Requirements

• Enrolled in a Bachelor ’ s or Master’s program in Robotics, Computer Science, or a closely related field
• Practical experience in implementing environmental recognition, from sensor input to human-readable output, in an academic (e.g., university project) or professional (e.g., work-study position) setting.
• Know-How in the field of Computer Vision
• Know-How in Machine/Deep Learning
• Know-How in C++
• Know-How in ROS2
• Know-How in Containerization

Nice to Have

• Know-How in the field of Audio processing
• Know-How in CUDA
• Know-How in Isaac ROS
• Know-How in using ONNX, TensorRT and Triton
• Know-How in shrinking Deep Learning models (quantization, pruning, compilation, ...)
• Know-How in Software Engineering
• Know-How in CI/CD

Why SWARM Biotactics

• We’re well-funded and expanding fast in a space no one else is in .
• You’ll join a world-class team working on the frontier of science and engineering .
• Your work will directly shape the foundation of biorobotic systems that redefine the interface between life and technology.

Join the Swarm

If you’re ready to create the impossible — hands-on, every day — we want to meet you. Feel free to apply directly here via LinkedIn or send your application materials to fabian.schwickert@swarm-biotactics.com. We look forward to hearing from you!