About the job
NVIDIA is redefining what is possible with AI, and we are building the next generation of relational deep learning for enterprise data. Our platform learns directly from the structure and relationships inside relational databases and heterogeneous graphs, unifying GPU‑accelerated graph analytics and graph machine learning in a single stack. It powers high‑impact workloads such as fraud detection and recommender systems and helps customers unlock more value from their data. We care deeply about turning frontier AI research into reliable products that teams can trust every day, and we are excited to add a Principal Technical Program Manager who will help us do exactly that!
Responsibilities
Lead the relational deep learning program from research to production, connecting ML researchers, infrastructure, and platform teams.
Deliver task‑specific models for domains such as fraud detection and recommender systems, moving from problem definition through training, benchmarking, and hand‑off.
Coordinate with infrastructure, systems, and platform groups to align compute capacity, training and serving environments, and platform features.
Guide release management for both the platform and models, including experiment‑to‑production hand‑offs, versioning, compatibility, model cards, benchmarks, and safety approvals.
Maintain the operating rhythm for the program, leading planning, reviews, risk and dependency tracking, and decision forums across partners.
Define and track program health metrics such as model quality, training speed, evaluation coverage, and time‑to‑release.
Qualifications
Minimum
Bachelor’s degree in Computer Science, Engineering, or a related technical field, or equivalent experience.
15+ years of experience in technical program management, engineering, or data/ML delivery, including significant time in ML/AI or large‑scale data environments.
Experience leading complex, multi‑stakeholder programs end to end in research and engineering organizations, with evolving requirements and clear delivery timelines.
Comfort working with ML researchers, interpreting model and evaluation results, and making decisions about training pipelines, data, and infrastructure trade‑offs.
Ability to build operating rhythms from scratch, influence without formal authority, and communicate clearly with both highly technical teams and senior executives.
Familiarity with modern program‑management practices (for example, Agile, roadmapping, risk and dependency management) and the judgment to use them effectively in fast‑moving research settings.
A hands‑on, builder mindset that includes creating automation and tooling, using AI in daily work, and applying AI to streamline program operations, status reporting, risk detection, and release workflows.
Preferred
Delivering graph ML, recommender, or fraud‑detection systems into production, or shipping ML platforms and frameworks that other teams build on.
Working with graph machine learning, GNNs, relational or tabular data, graph analytics libraries such as cuGraph, and the modern data stack (warehouses, feature stores, and data pipelines).
Running programs that span platform and infrastructure teams and model and research teams, including GPU and compute capacity planning for training and serving.