2 months ago
Responsibilities
- Own the technical direction, architecture, interfaces, and engineering standards for the MLOps platform.
- Design and operate dataset versioning, lineage, splits, reproducibility, and labeling-integration workflows.
- Build and operate a model registry containing versioned artifacts, metadata, evaluation results, lineage, and approval workflows.
- Define offline benchmarks, simulation rollouts, policy-gating harnesses, and metrics used to qualify models.
- Own the path from registered models to inference on Apollo, including packaging, on-robot versioning, rollback, and observability.
- Coordinate the deploy-and-telemetry connection with Connect and Data Platform.
- Lead technical projects from architecture through implementation, validation, and iteration.
- Mentor MLOps engineers through code reviews, design reviews, and direct collaboration.
- Establish platform contracts and standards across MLOps, Autonomy, Data Platform, TeleOp, and Training Infrastructure.
Requirements
- Deep proficiency in Python and at least one systems-level language: Go, Rust, or C++.
- 8+ years of professional software engineering experience in ML platforms or related infrastructure, or 4+ years of hands-on experience owning an MLOps platform that shipped models to production.
- Proven experience owning and delivering an end-to-end production MLOps platform covering datasets, experiment tracking, model registry, evaluation, and serving.
- Experience with dataset versioning using DVC, LakeFS, Delta, or an equivalent; experiment tracking using MLflow, W&B, or Determined; model registries; and policy serving.
- Strong experience designing service-oriented systems on Kubernetes and working across platform APIs and compute infrastructure.
- Experience defining ML model evaluation and qualification frameworks where regressions have high costs, such as robotics, safety-critical, or customer-facing production systems.
- Demonstrated ability to lead technical projects, influence cross-functional teams, establish adopted standards, and mentor engineers.
- Proficiency with AWS, GCP, or Azure, Docker, Git, and modern CI/CD workflows.
- Preferred experience deploying models to edge or embedded targets, including on-device inference, ONNX Runtime, TensorRT, or robot fleets.
- Preferred experience with reinforcement learning training and evaluation infrastructure, embodied agents, rollout workers, replay buffers, and simulation-evaluation harnesses.
- Familiarity with humanoid robotics, dexterous manipulation, teleoperation data, policy gating, shadow deployments, or staged autonomy rollouts.
- Experience with IsaacSim, MuJoCo, or equivalent simulation-in-the-loop evaluation.
- Open-source contributions to MLOps tooling such as MLflow, BentoML, KServe, or Ray Serve.
- A bachelor’s degree in Computer Science, Machine Learning, or a related technical field is considered; a master’s degree is preferred.
Benefits
- Direct-hire position.
- Equal employment opportunity and anti-discrimination protections are provided.
- The role requires prolonged desk and computer work, occasional lifting of up to 15 pounds, and the ability to read screens and communicate by hearing and speech.
Tech Stack
Categories
About Apptronik
Apptronik designs and manufactures AI-enabled humanoid robots, led by its Apollo platform, to automate tasks alongside people in manufacturing and logistics, with potential use in healthcare and homes. The privately held company is headquartered in Austin, Texas, and grew out of the University of Texas at Austin’s Human Centered Robotics Lab; its team previously worked on NASA’s Valkyrie robot. Apptronik’s business model centers on selling robots and the software and services needed to deploy them in industrial environments.