SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
The Exploration Company is building innovative aerospace technologies for space transportation. We are seeking a hands-on AI and Computer Vision Engineer to develop perception capabilities for autonomous close-proximity operations in space.
You will own the full pipeline for building deep-learning models that estimate the relative position and orientation of non-cooperative spacecraft from camera imagery. This includes designing and training 6-DoF pose estimation models, managing datasets (synthetic, lab, and orbital), implementing domain adaptation techniques, and optimizing models for flight-representative compute constraints.
Key responsibilities include:
- Designing, training, and evaluating deep-learning models for spacecraft pose estimation
- Owning the complete training pipeline: dataset generation, augmentation, domain adaptation, experiment tracking, and reproducibility
- Optimizing models for embedded and neuromorphic hardware through knowledge distillation, pruning, quantization, and quantization-aware training
- Porting and evaluating models on constrained hardware, characterizing accuracy versus energy trade-offs
- Building explainability and uncertainty quantification into the pipeline for failure mode debugging and certification readiness
- Exploring privacy-preserving and distributed training approaches for collaborative model improvement
- Prototyping self-supervised refinement methods for in-flight model adaptation
- Defining requirements, test scenarios, and validation criteria with guidance and flight operations teams
- Running validation campaigns on hardware-in-the-loop testbeds
- Managing training compute across cloud GPU and internal HPC resources
This is a builder role where you will write training code, run experiments, deploy models to hardware, and own the results with minimal daily direction but with review support.
REQUIREMENTS:
- MSc or PhD in computer science, electrical engineering, robotics, aerospace, physics, or comparable field with strong machine learning focus
- 3+ years building and shipping deep-learning computer vision systems (object detection, keypoint detection, pose estimation, 3D perception); PhD work counts
- Demonstrable experience optimizing models for constrained targets: quantization, distillation, latency optimization, deployment on embedded or accelerator hardware
- Experience training on synthetic data and managing sim-to-real gap
- Hands-on lab work: cameras, calibration, test setups, data collection and annotation
- Strong Python and PyTorch (or JAX/TensorFlow); clean, version-controlled, reproducible code
- Solid classical computer vision and 3D geometry: camera models, intrinsics/extrinsics, distortion, PnP, RANSAC, coordinate frames
- Comfortable with Linux, Git, containers, and running large training jobs on GPU clusters or cloud
- Genuinely hands-on and self-directed; able to carry work independently with review support
- Clear technical communication in reports and reviews
- Working proficiency in English; German is a plus
Nice to have: familiarity with spacecraft rendezvous/docking/vision-based navigation, SPEED/SPEED+ benchmarks, ESA pose estimation challenges, spiking neural networks, neuromorphic hardware, event-based cameras, federated learning, differential privacy, rendering/synthetic data generation (Blender, Unreal Engine, Isaac Sim), space or safety-critical software assurance experience, or publicly funded R&D project experience.
