AI and Computer Vision Engineer

The Exploration Company
Munich
Workplace: OnsiteFull timeFunction: Data Science & Machine LearningExperience: 3+ yearsEducation: mastersSkills: ["Self-directed","Technical communication","Reproducibility","Hands-on"]

Build deep-learning perception models for autonomous close-proximity space operations, estimating 6-DoF pose from camera imagery trained on synthetic and lab data. Own the end-to-end training pipeline—from dataset generation and sim-to-real domain adaptation to experiment tracking—then optimize for flight-representative compute with benchmarking. Port and evaluate models on embedded and neuromorphic hardware, add uncertainty/explainability, and validate via hardware-in-the-loop campaigns.

Loading

Loading job details...

Preparing the role view and application actions.

FursaFursa
The Exploration Company
The Exploration Company
3 days ago

AI and Computer Vision Engineer

✓ Verified Job

Canonical indexed version, validated from employer's careers page.

Source: Company careers pageValidated by: Fursa AI
Last checked: 11 hours agoStatus: Live

Job Summary

Build deep-learning perception models for autonomous close-proximity space operations, estimating 6-DoF pose from camera imagery trained on synthetic and lab data. Own the end-to-end training pipeline—from dataset generation and sim-to-real domain adaptation to experiment tracking—then optimize for flight-representative compute with benchmarking. Port and evaluate models on embedded and neuromorphic hardware, add uncertainty/explainability, and validate via hardware-in-the-loop campaigns.
Location: Munich
Workplace: Onsite
Employment Type: Full time
Job Function: Data Science & Machine Learning
Seniority: Mid level

Key Responsibilities

  • •Design, train, and evaluate deep-learning models for 6-DoF pose estimation of non-cooperative spacecraft from camera images.
  • •Own the full training pipeline, including dataset generation/management, augmentation, sim-to-real/domain adaptation, and experiment tracking for reproducibility.
  • •Optimize models for flight-representative compute using distillation, pruning, quantization, and quantization-aware training; benchmark latency, memory, and power against onboard constraints.
  • •Port and evaluate models on embedded and neuromorphic hardware, characterizing accuracy versus energy trade-offs and implementing explainability/uncertainty to debug failure modes.
  • •Define validation requirements and test scenarios with GNC/FPO, run hardware-in-the-loop validation campaigns, and manage training compute across cloud GPUs and internal HPC.

Pay and Benefits

Perks:Relocation

Key Requirements

  • •3+ years building and shipping deep-learning computer vision systems (object/keypoint/pose estimation or 3D perception), with PhD work counting.
  • •Experience taking a model from research prototype to constrained targets using quantization, distillation, latency optimization, and deployment on embedded/accelerator hardware.
  • •Experience training on synthetic data and addressing the sim-to-real gap.
  • •Hands-on lab work with cameras, calibration, test setups, and collecting/annotating data.
  • •Degree (MSc or PhD) in computer science, electrical engineering, robotics, aerospace, physics, or a comparable field with strong ML focus.
Experience:3+ years
Education:Master's
Skills:Self-directedTechnical communicationReproducibilityHands-on
Languages:EnglishGerman
Tech Stack:PythonPyTorchJAXTensorFlowLinuxGitContainersQuantizationPruningKnowledge distillationQuantization-aware trainingPnPRANSACBlenderUnreal EngineIsaac Sim

Eligibility

Work Authorization:Sponsorship available.

Company Brief

The Exploration Company
Designs and develops reusable crewed and cargo spacecraft for low Earth orbit missions. The company builds orbital transfer and return vehicles to enable routine transportation, research, and commercial services in space.
Industry: Space Technology
Company Size: Medium (51 to 250 employees)
Growth: Growth Stage Startup
Headquarters: Munich, Germany
Founded: 2018
WebsiteLinkedIn