Software Engineer Intern: ML Ops

Apptronik is seeking a Software Engineering Intern to join our ML Ops team for a 12-week summer project. In this role, you will help build a functional testing pipeline for our full simulation stack — the connective tissue that lets our simulation, controls, and learning teams iterate on hardware models, simulator versions, and ML controllers with confidence.

You will take ownership of a specific project: building a pipeline that runs defined policies against the simulator to validate changes across four dimensions — robot models (URDF/MJCF/USD), simulator code, controllers (open-source VLA, RL policies, VLA + RL action heads), and task scenarios. As a stretch objective, the same pipeline may be extended to generate synthetic training data from a “producer” model for use by a downstream “consumer” model. You will work alongside the simulation engineering and data platform teams, and contribute to infrastructure that directly accelerates how Apptronik trains and validates humanoid behaviors.

• Pipeline Development: Build and maintain a containerized functional testing pipeline that executes policies against the simulator in a reproducible, automatable fashion.
• Functional Test Harness: Implement test scaffolding that detects regressions across changes to robot models (URDF/MJCF/USD), simulator versions, and controller iterations, with baseline comparison and clear pass/fail signals (success rate, object drops, collisions).
• Synthetic Data Generation (stretch): Build a producer/consumer data generation flow that captures rollouts from one model and packages them into training-ready datasets for downstream consumers.
• Integration with Existing Infra: Integrate the pipeline with the team’s existing artifact storage (S3/MinIO), data lake formats (MCAP), and Kubernetes-based execution environment.
• Documentation & Handoff: Produce design docs, runbooks, and example configurations so the pipeline can be adopted by simulation, controls, and learning teams after the internship.

• Proficiency in Python: Demonstrated ability to write clean, tested, maintainable code for data pipelines, automation, and ML tooling.
• Linux & Development Tools: Comfortable in a Linux environment; competence with Git, Docker, and modern Python tooling (pytest, uv/poetry, type hints).
• Testing & Automation: Experience writing automated tests (unit, integration) and reasoning about determinism, flakiness, and reproducibility.
• ML / Data Pipeline Exposure: Familiarity with at least one ML framework (PyTorch, TensorFlow) and experience moving data between training, evaluation, and storage stages.
• CI/CD or Container Orchestration: Exposure to one or more of: GitHub Actions / GitLab CI / Bitbucket Pipelines, Kubernetes, Helm, ArgoCD. Deep expertise not required; willingness to learn the team’s stack is.
• Simulation Familiarity (preferred): Prior exposure to robotic or physics simulators (Isaac Sim / IsaacLab, MuJoCo, Gazebo, or comparable) and to robot description formats (URDF / MJCF / USD).
• Reinforcement Learning Exposure (preferred): Familiarity with RL training loops, policy rollouts, or vision-language-action (VLA) models.

• Current enrollment in a Bachelor’s or Master’s degree program in Computer Science, Electrical Engineering, Robotics, or a related field.
• Experience with projects involving simulation, ML model training, data pipelines, or developer tooling is ideal.

• Prolonged periods of sitting at a desk and working on a computer.
• Must be able to lift 15 pounds at times.
• Vision to read printed materials and a computer screen.
• Hearing and speech to communicate.