The opportunity
Software Operations is where data quality meets real-world impact. Our team reviews annotation outputs, tests the internal tools our engineers rely on, and surfaces the kinds of defects and patterns that, left unchecked, quietly degrade the systems powering our vehicles.
What you'll do
Review annotated datasets and labeling outputs for quality and consistency: across robotics and autonomy workflows.
Apply defined QC rubrics to evaluate outputs as pass, fail, or needs review,: documenting specific defect types clearly: incorrect output, regression, edge-case failure, data inconsistency, or tool behavior deviation.
Maintain a clean, complete review log for every item assessed. Every assessment is documented. No gaps.
Test internal tools across standard workflows and exception scenarios as directed by the Operations Lead.
Reproduce reported bugs reliably and validate fixes with clear supporting: documentation including steps taken, inputs used, expected versus actual behavior.
Track test coverage across tools, workflows, and edge cases, and flag discrepancies before each cycle begins.
What they're looking for
- Submit completed QC and testing logs on the agreed cadence, flagging urgent: or high-severity defects the same day, without waiting for the regular report.
- Write defect notes that are specific, reproducible, and immediately: actionable, not vague flags, but clear descriptions of what failed and why it matters.
- Escalate any case where the rubric or test plan is ambiguous or doesn't cover: the scenario. When you're unsure, you ask rather than guess.
- Flag recurring defects, systemic issues, or documentation gaps when you: notice patterns across multiple items. Your perspective across the full volume of work is something no one else has.
