Senior Manager, Training Data

NVIDIA•Santa Clara, CA

12d•$240,000 - $379,500

About The Position

NVIDIA is seeking a senior product leader to define the strategy and build platforms that deliver high-quality training data for its foundation models. The role begins with web and code data for Nemotron and focuses on helping research teams discover valuable open-web sources, improve data quality and diversity, and transform raw content into datasets that advance model capabilities. As a founding product leadership role, it partners with an engineering lead and works across research, data acquisition, data platforms, and legal teams to establish roadmaps and investment priorities. The position connects acquisition and curation decisions to measurable training outcomes while balancing quality, scale, delivery time, and cost, and may extend the platform to additional domains and modalities as research needs evolve. The role offers the opportunity to shape a foundational capability for NVIDIA AI models and work directly with researchers and engineers to improve what the models can learn. The successful candidate will bring technical curiosity, product judgment, and a record of delivering ambitious data platforms. The base salary range is USD 240,000–379,500, determined by location, experience, and pay of employees in similar positions. The role is also eligible for equity and benefits. Applications will be accepted at least until November 1, 2026. This posting is for an existing vacancy. NVIDIA uses AI tools in its recruiting processes and is an equal opportunity employer committed to an inclusive work environment.

Requirements

BS or MS in Computer Science, Engineering, or a related technical field, or equivalent experience.
8+ years overall in product management or a related technical role, including 5+ years on data platforms, search/indexing, or ML infrastructure.
4+ years of direct people-management experience.
Track record of delivering data systems at scale and improving quality, reliability, processing time, adoption, or cost; experience building a new platform or taking a technically complex product from concept to production.
Strong technical understanding of distributed data pipelines, storage, indexing or retrieval, and data processing; ability to engage in architecture and capacity discussions and make informed product tradeoffs with engineers.
Experience working closely with research or ML teams to turn evolving needs into datasets or platform capabilities; judgment in experiment design, baseline selection, and distinguishing data quality proxies from proven model improvement.
Strong analytical skills and ability to inspect data and performance metrics using SQL, Python, or equivalent tools; comfort making decisions with incomplete information and refining priorities as evidence develops.
Clear written and verbal communication, with proven ability to influence engineering, research, legal, and commercial stakeholders and secure decisions on scope, resources, and investment.

Nice To Haves

Direct experience with foundation-model training data, large-scale web crawling, source discovery, HTML parsing, or corpus curation.
Experience evaluating data suppliers, integrating publisher content, defining licensing requirements, or optimizing data acquisition economics.
Background with code, mathematical, multilingual, or multimodal datasets, semantic discovery, and quality ranking systems.
Experience establishing a product function, mentoring product managers, and shaping teams in partnership with engineering leaders.
Familiarity with synthetic data, reinforcement-learning environments, or post-training data workflows.

Responsibilities

Own the product vision, strategy, and roadmap for training data acquisition and preparation; translate research priorities and model capability gaps into source requirements, platform investments, and dataset acceptance criteria aligned with training schedules.
Define requirements for source discovery, URL indexing, predictive quality and diversity ranking, crawl scheduling, distributed fetching, and ingestion; partner with engineering on throughput, reliability, storage, and capacity planning as workloads scale.
Drive improvements in parsing and curation, including HTML extraction, browser-rendered content, code and math preservation, deduplication, quality filtering, and dataset composition; identify lost valuable content and increase usable, differentiated data yield.
Lead build, buy, and partner decisions across internal crawling, commercial datasets, and licensed content; establish comparable internal and vendor pilots measuring incremental coverage, accepted data yield, delivery reliability, and total cost against existing corpora.
Partner with researchers on controlled training experiments and ablations to measure data impact on model capabilities; account for benchmark contamination, dataset composition, and regressions, and use results to guide source selection and acquisition investment.
Work with legal, privacy, policy, and security teams to translate approved requirements into scalable controls for source eligibility, access restrictions, opt-outs, provenance, and permitted use; define auditable dataset acceptance and release processes.
Build researcher-facing capabilities for dataset discovery, metadata, versioning, lineage, and quality reporting; enable reproducible dataset releases and reuse, measure adoption, and reduce time from research request to training-ready data.
Align teams on priorities, dependencies, and ownership; shape hiring and resource plans with engineering leadership, communicate investment tradeoffs to senior leaders, and establish shared metrics for model impact, data coverage, pipeline latency, cost, and delivery against training cutoffs.