About The Position
NVIDIA is building the future of autonomous driving, and evaluation is how we know the drive is getting better. Our AV Evaluation team owns the metrics, golden datasets, and closed-loop evaluation workflows that decide what ships in NVIDIA's self-driving stack — every model change runs through us. We are bringing frontier vision-language models and agentic techniques to evaluation at scale, and the early results are changing how our organization develops AI drivers! We are looking for a senior engineer to own the engine that makes it all trustworthy: the eval flywheel. This is a hands-on technical leadership role — the flywheel's architecture, quality, and adoption are owned end to end by this person, in code and in the room, not from a strategy document.
Requirements
- A track record of independent execution and technical leadership: finding the highest-leverage problem, driving it across team boundaries, and delivering without waiting to be asked.
- Clear, proactive communication with engineers and senior leaders alike, at a fast pace.
- BS or MS in Computer Science, Robotics, or a related field (or equivalent experience).
- 12+ years building software, with significant time in autonomous vehicles, robotics, or large-scale ML systems.
- Deep experience evaluating ML or robotic systems: metric design, ground-truth and golden dataset curation, precision/recall methodology, and the data pipelines behind them.
- Strong Python and data engineering skills for production-scale pipelines.
Nice To Haves
- Experience building an evaluation flywheel before — dataset curation, metric measurement, developer tooling — and the story of how it changed model development velocity.
- Closed-loop simulation evaluation for autonomous driving, and the realism questions that come with it.
- Applying LLMs or VLMs to evaluation, or productionization the infrastructure behind them.
- A history of earning trust for metrics across skeptical partner teams.
Responsibilities
- Owning the eval flywheel's strategy and architecture: how road and simulation driving data becomes curated golden datasets, how metrics are measured against them (precision/recall), and how those results earn lasting trust with the teams that depend on them.
- Setting the standard for evaluation quality: golden dataset curation, versioning, and health; metric performance measurement; and release processes that keep results dependable as the system evolves.
- Building the tooling that helps our metric developers iterate quickly: self-serve dataset pipelines, metric performance measurement, and quality reporting used every day by the team and our partners.
- Partnering with senior engineers and leaders across test engineering, behavior planning, and infrastructure — setting expectations, working through trade-offs, and being the voice of evaluation quality in cross-team decisions.
- Working directly with AI model developers so evaluation iteration speed becomes an advantage for the whole program, including our push into learned, VLM-based evaluation.
Benefits
- equity
- benefits
Stand Out From the Crowd
Upload your resume and get instant feedback on how well it matches this job.
What This Job Offers
Job Type
Full-time
Career Level
Senior
Education Level
Associate degree
