Senior Data Scientist, Data Quality and AI Evaluation

NVIDIA•Santa Clara, CA

About The Position

At NVIDIA, this role supports development of an AI-native autonomous driving architecture that combines classical safety systems, foundation models, and scalable AI systems. The Senior Data Scientist will evaluate data quality and auto-labeling pipelines, identify where they succeed, fail, or regress, and work with engineering and labeling operations teams to improve them. The role involves building human-verified golden datasets, analyzing large-scale sensor, camera, and annotation data, and developing methods to cluster failures and select data for human correction. The position also establishes statistically sound audits, experiments, and release acceptance criteria, and communicates findings through dashboards and recommendations. It partners with labeling operations to improve annotation guidance, reviewer consistency, and the balance among quality, efficiency, and cost.

Requirements

Strong foundations in statistics, experimental design, sampling, hypothesis testing, and uncertainty estimation, with practical experience evaluating machine learning systems and data quality.
Experience with AI/LLM evaluation and proficiency in Python and SQL, including analysis of large datasets and development of reproducible analytical workflows.
Understanding of precision, recall, class imbalance, selection bias, data leakage, and evaluation across significant scenario groups.
Ability to distinguish reference-label defects from model errors and turn ambiguous quality concerns into measurable questions and actionable recommendations.
Strong communication and teamwork skills, with experience working across engineering, research, product, or operations teams.
12+ years of industry experience, with a bachelor’s degree or proven track record for a master’s degree in statistics, mathematics, computer science, engineering, or a related quantitative field, or equivalent practical experience.

Nice To Haves

Background in autonomous driving, robotics, or computer vision, particularly with camera and LiDAR data, 2D/3D annotations, multi-object tracking, occupancy, or temporal and cross-sensor consistency.
Experience designing human-in-the-loop labeling and audit workflows, measuring annotator agreement, or maintaining versioned benchmarks with protected evaluation sets.

Responsibilities

Define quality metrics and evaluation methods for identifying successes, failures, and regressions in auto-labeling pipelines.
Build and maintain human-verified golden datasets representing driving conditions, corner cases, historical failures, and new requirements.
Analyze large-scale sensor, camera, and annotation data to find root causes, systematic errors, and improvement priorities in collaboration with perception and auto-labeling engineers.
Develop automated methods for failure clustering, similar-case discovery, and targeted selection of data for human correction.
Build statistically sound audits and experiments, establish release acceptance criteria, and communicate findings through dashboards and recommendations.
Partner with labeling operations to improve annotation guidelines, reviewer consistency, and the balance of quality, efficiency, and cost.