AI Infrastructure & Robotics DevOps Intern

📋 Job Overview

Robots don't learn by magic, they learn because something quietly spins up thousands of simulations at 2am, ships fresh models to the lab by breakfast, and keeps every machine in sync while nobody's watching. That \"something\" is what you'll be building with us.

You'll join a team that trains, tests, and deploys AI on real humanoid robots every single day, and you'll own the invisible infrastructure that makes all of it feel effortless. Expect a playground of GPUs, self-hosted runners, simulation clusters, and pipelines that stretch from a git push all the way onto a robot standing in our lab.

🏢 About neura robotics gmbh

Currently pursuing or recently completed a degree in Computer Science, Software Engineering, Electrical Engineering, or a related field.

🎯 The Role

CI/CD Engineering: Shape our pipelines across GitLab and GitHub Actions, matrix builds, smart caching, multi-stage workflows for robotics code and ML training jobs.
Infrastructure as Code: Tame our fleet of runners, GPU workstations, and lab machines with Ansible, so \"it works on my machine\" becomes \"it works on all of them\".
Python Packaging & Tooling: Turn internal libraries and controllers into clean, versioned pip-installable packages with uv, and help design how we host and consume them internally.
Containers Everywhere: Build and optimize Docker images (yes, including the fun DinD / DooD rabbit holes) for training, simulation, and deployment — fast, small, reproducible.
Simulation at Scale (Isaac Sim): Bring NVIDIA Isaac Sim / Isaac Lab into CI: headless simulation containers, scenario- and policy-level regression tests, and a safety net that catches bugs before they meet real hardware.
Deployment Pipelines: Ship validated artifacts, Python packages, Docker images, model checkpoints, sim assets, from CI all the way onto lab machines and physical robots, with rollbacks that actually work.
Nightly Runs: Own the heartbeat of the team: nightly training, evaluation, simulation, and integration runs. Schedule them, watch them, triage them, make them better.
Git & Repo Hygiene: Help us stay sane with submodules, LFS, code owners, protected branches, and no more 2GB accidental commits.

✅ Key Responsibilities

• Shape our pipelines across GitLab and GitHub Actions, matrix builds, smart caching, multi-stage workflows for robotics code and ML training jobs.
• Tame our fleet of runners, GPU workstations, and lab machines with Ansible, so \"it works on my machine\" becomes \"it works on all of them\".
• Turn internal libraries and controllers into clean, versioned pip-installable packages with uv, and help design how we host and consume them internally.
• Build and optimize Docker images (yes, including the fun DinD / DooD rabbit holes) for training, simulation, and deployment — fast, small, reproducible.
• Bring NVIDIA Isaac Sim / Isaac Lab into CI: headless simulation containers, scenario- and policy-level regression tests, and a safety net that catches bugs before they meet real hardware.
• Ship validated artifacts, Python packages, Docker images, model checkpoints, sim assets, from CI all the way onto lab machines and physical robots, with rollbacks that actually work.
• Own the heartbeat of the team: nightly training, evaluation, simulation, and integration runs. Schedule them, watch them, triage them, make them better.
• Help us stay sane with submodules, LFS, code owners, protected branches, and no more 2GB accidental commits.

📌 Required Qualifications

• Currently pursuing or recently completed a degree in Computer Science, Software Engineering, Electrical Engineering, or a related field.
• Solid proficiency in Python, plus comfort with Bash and Linux command-line tooling.
• Hands-on experience (from studies, side projects, or previous internships) with at least some of: Docker, GitLab CI, GitHub Actions, Ansible, uv/pip, Git submodules & LFS.
• Understanding of CI/CD concepts: caching, artifacts, runners, pipelines, reproducible builds.
• Curiosity about MLOps and how to keep large ML/robotics training pipelines healthy and cost-efficient.
• Strong problem-solving and debugging skills, you enjoy figuring out why a pipeline is red at 3am (figuratively).
• You care about developer experience and want to make life easier for the engineers around you.
• You have a good command of the English language and, ideally, also speak German.

⭐ Desirable Experience

• Exposure to NVIDIA Isaac Sim / Isaac Lab, ROS 2, headless GPU simulation in containers, or automated deployment to robots / edge devices.

🎁 Benefits

Competitive

🛂 Visa & Eligibility

Not specified