Мэтч & Сопровод
Для мэтча с этой вакансией нужен Plus
Описание вакансии
Текст:
TL;DR
AI and Computer Vision Engineer (AI/aerospace): Building perception models for autonomous close-proximity spacecraft operations, including 6-DoF pose estimation from camera images, with an accent on synthetic-data training, sim-to-real adaptation, and deployment on flight-representative hardware. Focus on optimizing and validating models under onboard latency, memory, power, uncertainty, and certification constraints.
Location: Munich, Germany
Company
The Exploration Company develops aerospace technologies for the future of space transportation.
What you will do
- Design, train, and evaluate deep-learning models for 6-DoF pose estimation of non-cooperative spacecraft.
- Own the end-to-end training pipeline, including dataset generation, augmentation, sim-to-real adaptation, experiment tracking, and reproducibility.
- Optimize and deploy models on embedded and neuromorphic hardware using distillation, pruning, quantization, and quantization-aware training.
- Build explainability, uncertainty estimation, self-supervised refinement, and privacy-preserving or distributed training capabilities.
- Define requirements, test scenarios, validation criteria, and camera-sensor characterization with GNC/FPO teams.
- Run hardware-in-the-loop validation campaigns and manage training workloads across cloud GPUs and internal HPC infrastructure.
Requirements
- MSc or PhD in computer science, electrical engineering, robotics, aerospace, physics, or a comparable field with a strong machine-learning focus.
- At least 3 years of experience building and shipping deep-learning computer-vision systems; PhD research in the field counts.
- Experience taking models from research prototypes to constrained targets, including quantization, distillation, latency optimization, and embedded or accelerator deployment.
- Strong Python and PyTorch skills, or equivalent JAX/TensorFlow experience, with clean, version-controlled, reproducible code.
- Knowledge of classical computer vision and 3D geometry, including camera models, calibration, PnP, RANSAC, and coordinate frames.
- Working proficiency in English is required; German is a plus.
Nice to have
- Experience with spacecraft rendezvous, docking, vision-based navigation, SPEED/SPEED+, or ESA pose-estimation challenges.
- Experience with spiking neural networks, neuromorphic hardware, or event-based cameras.
- Knowledge of federated learning, differential privacy, secure aggregation, or synthetic-data rendering tools.
- Experience with space or safety-critical software assurance or publicly funded R&D projects.
Culture & Benefits
- Fast decision-making within an Agile working environment.
- Open and collaborative communication focused on risks, obstacles, and shared problem-solving.
- Relocation assistance is provided for candidates willing to relocate, including visa sponsorship where applicable.
- Employment offers are subject to a background check depending on the country, role, and project.
Будьте осторожны: если работодатель просит войти в их систему, используя iCloud/Google, прислать код/пароль, запустить код/ПО, не делайте этого - это мошенники. Обязательно жмите "Пожаловаться" или пишите в поддержку. Подробнее в гайде →