Staff Software Engineer, ML

Apptronik•Sunnyvale, CA

About The Position

Apptronik is seeking a Staff MLOps Engineer to own the technical direction of its MLOps platform, connecting teleoperation data collection with deployed autonomy on its Apollo humanoid robot. The role sets architecture for the platform layer above the training cluster, including dataset lifecycle, experiment tracking, model registry, evaluation harnesses, and the serving and packaging path for delivering trained policies to robots. This is a hands-on technical leadership position, not a management role. The engineer will contribute directly while mentoring colleagues and leading by influence across MLOps, Autonomy, Data Platform, and TeleOp. The role partners closely with the Training Infrastructure engineer, who owns the underlying cluster layer, and helps turn research code into a repeatable, auditable pipeline from data to deployed model.

Requirements

Deep proficiency in Python and at least one systems-level language: Go, Rust, or C++.
Demonstrated ability to make and defend architectural tradeoffs in production ML platforms.
Proven experience owning and delivering an MLOps platform end-to-end, including dataset lifecycle, experiment tracking, model registry, evaluation, and serving, at a company that ships models to production.
Expertise across the model lifecycle, including dataset versioning, experiment tracking, model registry, and policy serving.
Strong background designing service-oriented systems on Kubernetes and understanding the contract between platform APIs and underlying compute infrastructure.
Experience defining evaluation and qualification frameworks for ML models where regressions are costly, such as robotics, safety-critical, or production-customer-facing environments.
Experience leading technical projects end-to-end, including architecture, implementation, validation, and iteration.
Demonstrated ability to lead by influence across teams, establish standards other engineers adopt, and mentor engineers.
Proficiency with cloud infrastructure such as AWS, GCP, or Azure, Docker, Git, and modern CI/CD workflows.
Master's degree in Computer Science, Machine Learning, or a related technical field preferred; a bachelor's degree may be considered with exceptional experience.
At least 8 years of professional software engineering experience in ML platforms or related infrastructure, or at least 4 years of direct, hands-on experience owning an MLOps platform that shipped models to production.

Nice To Haves

Experience deploying ML models to edge or embedded targets, including on-device inference, ONNX Runtime, TensorRT, or robot fleets.
Experience with RL training and evaluation infrastructure for embodied agents, including rollout workers, replay buffers, or simulation-evaluation harnesses.
Familiarity with humanoid robotics, dexterous manipulation, or teleoperation data domains.
Experience with simulation-in-the-loop evaluation, such as IsaacSim, MuJoCo, or equivalent.
Familiarity with policy gating, shadow deployments, or staged rollout strategies for autonomy.
Open-source contributions to MLOps platform tooling such as MLflow, BentoML, KServe, or Ray Serve.

Responsibilities

Own the technical direction and architecture of the MLOps platform, including subsystem interfaces, architecture decisions, and engineering standards for dataset, experiment, and model workflows.
Serve as the primary technical point of contact for Autonomy, Data Platform, and TeleOp on model lifecycle and platform contracts.
Design and operate the dataset lifecycle, including versioning, lineage, splits, and labeling-integration handoff.
Ensure trained models are traceable to the exact data and code used to produce them.
Build and operate a model registry with versioned artifacts, metadata, evaluation results, lineage, and approval workflows.
Define the promotion path for models from trained to qualified to deployed on a robot.
Define offline benchmarks, simulation rollouts, and policy-gating harnesses that models must pass before deployment to Apollo.
Develop metrics frameworks used by the autonomy team to gate releases.
Own the path from registered model to on-robot inference, including packaging, on-robot versioning, rollback, and observability of deployed policy behavior.
Coordinate with Connect and Data Platform on the deployment and telemetry connection back from the fleet.
Mentor mid-level and senior MLOps engineers through code reviews, design reviews, and direct collaboration.
Partner with the Training Infrastructure engineer on the cluster/platform contract and influence Autonomy research workflows to standardize on platform primitives.