Engineering Manager - Autonomy Subsystem Evaluation

Job Description About the Organization GM has more than 500,000 Super Cruise–equipped vehicles on the road and over 700 million hands-free miles driven. We're now building the next generation of autonomous driving, starting with eyes-off driving on the highway and growing toward point-to-point autonomy. The goal is autonomy that is safe, reliable, comfortable, and scalable across GM's vehicle portfolio. Autonomy Evaluation helps the AV program figure out where the driver still falls short. Every mile our vehicles drive, on the road or in simulation, carries that signal. We turn it into answers: what is limiting the system, how big each problem is, and what fixing it is worth. Those answers steer what GM's autonomy teams build next and inform software release decisions. Subsystem Evaluation is one of four teams in Eval Applications, the part of Autonomy Evaluation that turns evaluation into decisions. Each autonomy team measures its own component, usually by accuracy against labels. We answer the question those metrics can't: how much does each subsystem's error actually affect driving? We build metrics for planning, perception, prediction, and localization, including learned components, that reflect their effect on the drive. We also trace driving failures back to the subsystem or interaction that caused them. The Role As Manager of Subsystem Evaluation, you'll lead a team of engineers and data scientists and own its roadmap, execution, and staffing. The work is concrete. One example is showing which perception errors change what the car does and which don't. Another is ranking the subsystem failures behind a drop in driving quality, so the right team fixes the right thing first. You'll work with the autonomy teams that use these metrics, with the Driver Performance and Scenario Classification teams, and with the principal and staff engineers who lead evaluation methodology. You'll report to the Senior Manager of Eval Applications. What You'll Do • Set the roadmap and quarterly plans for subsystem evaluation, tied to the AV program's milestones and release needs. Decide which subsystems and questions come first, and what to stop. • Build metrics that weight subsystem errors by their effect on downstream decisions and driving, not by accuracy alone. • Trace bad driving in simulation and road data back to the subsystem or interaction that caused it, and measure how subsystem errors add up to driving outcomes. • Give autonomy teams and release reviews a fast, trustworthy answer to whether a software or model change helped or hurt driving, including regression detection and version-to-version comparison. • Hold the metrics themselves to a high bar. Check them against driving outcomes and human judgment, state their confidence and limits, and retire the ones that don't hold up. • Apply VLMs and LLMs effectively where applicable in the development/trigaging/etc processes.   • Turn results into clear findings for autonomy teams and leadership: what changed, how confident we are, and what to fix first. Leadership Responsibilities • Hire, develop, and retain engineers and data scientists, with clear expectations, regular feedback, and fair performance management. • Set clear ownership and steady operating rhythms. Keep the technical bar high through design reviews, code review, testing, and statistically sound analysis. • Balance near-term program requests against longer-term investments in metrics and methods. • Build trust with autonomy teams so they act on our metrics. Agree with them on where their component metrics end and ours begin, and resolve disagreements over priorities or conflicting results. Your Skills & Abilities (Required Qualifications)   • 8+ years in autonomous systems, robotics, ML evaluation, or appli