AI/ML Engineer

Infinite Orbits PTE Ltd.; Toulouse; Hybride; EUR 65 000 - 95 000; Plein temps

Résumé du poste

Infinite Orbits PTE Ltd. in Toulouse is seeking an AI/ML Engineer to design, train, and validate learning-driven models that operate onboard spacecraft, supporting perception, navigation, and decision-making in real mission environments.

This hands-on position requires building flight-ready ML components, integrating Python and C++ flight software, and collaborating with an international team across simulations, testing, and deployment of autonomous space systems.

Qualifications

• Strong foundation in machine learning, deep learning, reinforcement learning, or a related technical discipline.
• Practical experience developing, training, and evaluating ML models for perception, estimation, planning, or control systems.
• Proficiency with modern ML frameworks such as PyTorch or TensorFlow, and the ability to integrate models into Python and C++ systems.
• Experience building simulation tools or data generation pipelines to support ML development.
• Ability to design, debug, and scale training workflows, including dataset management, distributed training, and evaluation.
• Familiarity with deploying ML inference as part of larger autonomous or robotic software stacks.

Responsabilités

• Design, train, and validate AI/ML models supporting satellite autonomy, including rendezvous and proximity operations, on-orbit inspection, formation flying, and space domain awareness.
• Develop learning-based components for guidance, navigation, perception, and decision-making, such as pose estimation, anomaly detection, behavior recognition, and trajectory generation.
• Build LLM- and VLM-powered tools to enable intuitive satellite operations, including natural language command interfaces and visual scene understanding.
• Create and maintain synthetic data generation pipelines that model realistic orbital dynamics, sensor behavior, lighting conditions, and uncertainty, using domain randomization and sim-to-real techniques.
• Extend and improve simulation environments used for training, verification, stress testing, and adversarial evaluation of autonomy models.
• Optimize models for flight deployment, applying techniques such as quantization, pruning, batching, and real-time inference on limited onboard compute.
• Integrate ML components into flight software using well-structured Python and C++ interfaces, with strong safeguards and fallback behaviors.
• Participate in software-in-the-loop and hardware-in-the-loop testing, closing the feedback loop between simulation, sensors, control systems, and model improvements.
• Take full ownership of your models, from early experimentation and training pipelines to testing, review, deployment, and long-term maintenance.

Connaissances

ML fundamentals

Deep learning

Reinforcement learning

Python integration

Outils

PyTorch

TensorFlow

C++

Python

Simulation tools

Description du poste

Established in 2017, Infinite Orbits is a NewSpace company, based in Toulouse, pushing the boundaries of in-orbit services with cutting-edge spacecraft driven by its own Rendez-Vous solution, an advanced Autonomous Vision-Based Navigation Solution. IO has two generations of missions

"Orbit Guard" provides Inspection and Space Situational Awareness services in GEO. Our first, OG-1 was launched on 1 May 2023, the second is launching in 2025 and the third third launching in 2026.

"Endurance" is a docking service capable of extending the life of GEO telecom satellites, by taking over AOCS takeover for 5 years.

About the Role

We’re seeking an AI/ML Engineer to build and deploy learning-driven capabilities within our autonomous space systems. In this role, you’ll create models that directly operate in closed-loop autonomy, supporting perception, navigation, and decision-making in real mission environments. This is a hands-on position for engineers who want their models running on real spacecraft, not just evaluated in offline experiments.

Job Description

Design, train, and validate AI/ML models supporting satellite autonomy, including rendezvous and proximity operations, on-orbit inspection, formation flying, and space domain awareness.

Develop learning-based components for guidance, navigation, perception, and decision-making, such as pose estimation, anomaly detection, behavior recognition, and trajectory generation.

Build LLM- and VLM-powered tools to enable intuitive satellite operations, including natural language command interfaces and visual scene understanding.

Create and maintain synthetic data generation pipelines that model realistic orbital dynamics, sensor behavior, lighting conditions, and uncertainty, using domain randomization and sim-to-real techniques.

Extend and improve simulation environments used for training, verification, stress testing, and adversarial evaluation of autonomy models.

Optimize models for flight deployment, applying techniques such as quantization, pruning, batching, and real-time inference on limited onboard compute.

Integrate ML components into flight software using well-structured Python and C++ interfaces, with strong safeguards and fallback behaviors.

Participate in software-in-the-loop and hardware-in-the-loop testing, closing the feedback loop between simulation, sensors, control systems, and model improvements.

Take full ownership of your models, from early experimentation and training pipelines to testing, review, deployment, and long-term maintenance.

Requirements

Strong foundation in machine learning, deep learning, reinforcement learning, or a related technical discipline.

Practical experience developing, training, and evaluating ML models for perception, estimation, planning, or control systems.

Proficiency with modern ML frameworks such as PyTorch or TensorFlow, and the ability to integrate models into Python and C++ systems.

Experience building simulation tools or data generation pipelines to support ML development.

Ability to design, debug, and scale training workflows, including dataset management, distributed training, and evaluation.

Familiarity with deploying ML inference as part of larger autonomous or robotic software stacks.

A sense of ownership, adaptability in fast-paced environments, and clear communication around technical tradeoffs.

Nice to Have

Experience working with foundation models or large-scale pretrained architectures.

Knowledge of orbital mechanics, spacecraft dynamics, or space systems.

Experience combining learned models with classical estimation, guidance, or control algorithms.

Background in sim-to-real transfer, domain adaptation, or operational ML systems.

Exposure to embedded inference, GPU optimization, or real-time autonomous platforms.

Publications, open-source contributions, or research experience in ML, robotics, autonomy, or reinforcement learning.

What we offer

The opportunity to be part of an international team transforming the space industry.

A creative and innovative work environment where ideas turn into reality.