Fleet Telemetry Anomaly Detection

The brief

Orbital Loop's autonomous fleet generates several million telemetry events daily with no systematic anomaly detection. You will design and ship a detection pipeline, tune it against six months of historical incident data, and hand over an on-call runbook the operations team can actually use.

What you'll need

• Comfortable with streaming data pipelines
• Some exposure to time-series or anomaly detection methods

What you'll walk away with

• A production anomaly detection pipeline with a measured false-positive rate
• An on-call runbook adopted by the operations team