Senior Software Engineer, Autonomy Safety Metrics

Quick Facts

• Senior role focused on autonomy safety metrics through behavior evaluation

Description

You will build and scale the pipeline that produces weekly per-geofence safety estimates for autonomous driving behavior. The work converts statistical safety methodology—such as sampling, triage weights, mileage sharing, and projections—into tested, reproducible production code that fails loudly on bad data. You will create tooling to reduce manual weekly runs and collaborate with stakeholders to apply the methodology consistently in code and clearance reports, including adapting to new releases and new geofences.

Responsibilities

• Build and maintain Behavior Evaluation data pipelines on Databricks (PySpark/zSpark medallion architecture), plus EMR Serverless and Airflow
• Produce weekly per-geofence estimates
• Implement and validate statistical methods in reproducible code with data quality checks
• Develop tools to reduce manual effort in weekly runs
• Ensure cross-functional alignment so the methodology is implemented consistently in code and clearance reports
• Adapt the evaluation approach to new release methodologies and geofences

Requirements

• 7–9 years experience, or 5+ years with a strong data-platform track record
• Expert Python with production experience in PySpark/Spark and Pandas
• Experience designing scalable batch data pipelines (orchestration, Delta/lakehouse tables, idempotent backfills)
• Applied statistics knowledge (sampling, weighting, estimators, confidence intervals)
• Strong testing and CI habits for data code

Benefits

• Fast-paced, collaborative environment as the autonomy system expands to more cities and software releases
• Opportunity to improve speed, reliability, and auditability of autonomy safety metrics