Live sensor streams from autonomous systems at sea. Want to build the platform that makes all of that data trustworthy for ML?
A company in the LA area is building autonomous maritime systems. Their ML and perception teams live on sensor data, and they need a platform that lets them process live signals, replay historical recordings and know exactly which data shaped a model or an eval. You build that platform, spanning software infrastructure, dataset quality and close work with the teams building the autonomy itself.
What you'll own
Pipelines that ingest and process high-volume sensor and telemetry data for ML and analytics
Aligning and calibrating data across sensors, for near-real-time processing and repeatable historical reprocessing
Traceable datasets through versioning, metadata and lineage, so every experiment can be reproduced
Data-discovery interfaces and access patterns that help ML and perception teams find the right data fast
Measuring completeness and quality, digging into failures and keeping things reliable as volume grows
Turning shifting model needs into durable data contracts, together with ML, hardware and product
What you bring
You've built and run serious data platforms, ML data infrastructure or large-scale processing pipelines
Strong Python, with hands-on use of tools like PyArrow, Polars, Pandas or NumPy
Experience with sensor, time-series, audio, vision or similarly complex data
A solid grasp of dataset quality, reproducibility and how downstream users access data
You own technical decisions and work easily across software and ML teams
Bonus points
Sonar, maritime, robotics, autonomy or other sensor-rich environments
Orchestration, metadata or experiment-tracking tools
Backfills, labeling workflows, data APIs/SDKs or storage optimization at scale
What's in it for you
$200k–$240k base + equity
Full-time, Torrance, California
Data work that directly shapes how autonomous systems at sea see the world
Does this sound like you? Apply now. If you're a match, we'll be in touch within 3 working days.