Machine Learning Engineer

Senior Machine Learning Engineer role at Destinusgroup in Zürich, focusing on computer vision for autonomous flight in defense applications. Requires 5+ years in deep learning, strong C++ or Rust skills, and experience across the full ML lifecycle.

This role is part of Daedalean AG, the AI and autonomy hub of the Destinus Group, which develops safety-critical AI systems for aviation. It involves creating intelligent systems for autonomous flight technologies in defense, combining deep learning, simulation, and rigorous verification into a seamless pipeline.

Responsibilities

• Design, train, and optimize deep learning models for computer vision in dynamic, safety-critical environments
• Manage the full machine learning lifecycle, from data strategy and model architecture to deployment and performance monitoring
• Create robust evaluation pipelines to ensure models meet strict reliability and safety standards
• Participate in ML certification efforts, validating model behavior under regulatory frameworks
• Use transfer learning and simulation to extend training and testing beyond real-world data constraints
• Collaborate with software, systems, and flight teams to integrate ML models into real-world applications
• Continuously improve model performance across edge cases, environmental changes, and operational limits

Requirements

• Strong programming skills in C++ or Rust
• Master’s or PhD in computer science, physics, mathematics, or a related technical field
• At least 5 years of hands-on experience in deep learning for computer vision
• Experience across the full ML stack, including model design, training pipelines, and evaluation systems
• Solid understanding of data-centric AI, including dataset curation and augmentation
• Proven ability to solve complex research problems over long periods in academic or industrial settings

Nice to have

• Experience with simulation environments or synthetic data generation

About the company

The company emphasizes long-term thinking, rigorous validation, and ownership of real-world performance. It operates in challenging, ambiguous environments where problems are complex and meaningful, pushing beyond 'good enough' to achieve reliability, scalability, and certifiability.

• Values long-term thinking and rigorous validation
• Thrives on solving hard, ambiguous, and meaningful problems
• Committed to achieving reliability, scalability, and certifiability beyond basic performance