About this role
This role owns NVIDIA’s GenAI data strategy to fuel frontier foundation models via a “data flywheel.” You will lead end-to-end multi-lingual, multi-modal data acquisition and governance, drive failure analysis and dataset iteration, and oversee HITL and synthetic data methods to improve model alignment and safety.
Key Responsibilities
- Define and evolve the end-to-end roadmap for multi-lingual, multi-modal data acquisition
- Lead the data flywheel with model failure analysis and continuous data iteration
- Manage external data partners/vendors with qualification and acceptance criteria
- Own HITL strategy including RLHF and SFT workflows for safety and alignment
- Establish data quality and governance frameworks including de-duplication, versioning, bias detection, and ethical filtering
Technical Overview
The scope includes LLM/VLM architectures, training regimes, and alignment approaches such as RLHF and RLAIF. The director will manage multi-modal dataset sourcing, curation, synthetic generation for fine-tuning and evaluation, and establish governance frameworks (deduplication, versioning, bias detection, privacy/consent/transparency) while coordinating customer and ecosystem data priorities.
Ideal Candidate
The ideal candidate is an executive-level AI/ML data leader with 18+ years across product management or data operations in the AI/ML domain and 8+ years of direct people management experience. They deeply understand LLM/VLM training regimes and alignment methods such as RLHF and RLAIF, and they can build a comprehensive GenAI data flywheel strategy spanning data acquisition, curation, synthetic generation, and governance.
Must-Have Skills
Bachelor's or Master's degree in Computer ScienceAI/MLData Scienceor a related technical field (or equivalent experience)18+ overall years of experience in product management or data operations specifically within the AI/ML sector at a technology company8+ years direct people management experienceDeep understanding of LLM/VLM architecturestraining regimesand alignment methods like RLHF and RLAIF
Required Skills
data strategyend-to-end roadmapmulti-lingual data acquisitiondata flywheelmodel failure modesstrategic data acquisitiondata curationsynthetic generationgolden bench setHuman-in-the-Loop (HITL)RLHF (Reinforcement Learning from Human Feedback)SFT (Supervised Fine-Tuning)RLAIFLLM/VLM architecturesdata partnersdata governancede-duplicationversioningbias detectionethical filteringdata privacyconsenttransparencybenchmarksdataset release
Hard Skills
data strategyend-to-end roadmapmulti-lingual data acquisitiontextvisionaudiofoundation modelsdata flywheelmodel failure modes analysisdata acquisition logicdata curationsynthetic generationgolden bench setagentic capabilitiesdata partner and vendor managementqualification and acceptance criterialicensed assets compliancelegal standardsethical standardstechnical standardsHuman-in-the-Loop (HITL)RLHF (Reinforcement Learning from Human Feedback)SFT (Supervised Fine-Tuning)human-in-the-loop workflowssynthetic data streamfine-tuningevaluation techniquesdata governancede-duplicationversioningbias detectionethical filteringdata privacyconsenttransparencyLLM/VLM architecturestraining regimesalignment methodsRLAIFcustomer engagementSolutions Architectsenterprise customersbenchmarks for developer adoptiondataset releasedata quality frameworks
Soft Skills
leadershipstrategic thinkingcross-functional collaborationpartnering with research scientists and engineerspartnering with Solutions Architects and enterprise customerspeople managementcommunicationability to interpret model edge cases and failure modesstakeholder management
Keywords for Your Resume
Senior DirectorSenior Director - Gen AI Data StrategyGen AI Data Strategydata flywheelfoundation modelsdata strategyend-to-end roadmapmulti-lingual data acquisitiontext vision audiogolden bench setagentic capabilitiesmodel failure modesstrategic data acquisitiondata curationsynthetic generationHuman-in-the-Loop (HITL)RLHF (Reinforcement Learning from Human Feedback)SFT (Supervised Fine-Tuning)RLAIFLLM/VLM architecturestraining regimesdata partnersdata quality frameworksde-duplicationversioningbias detectiondata privacyconsenttransparencybenchmarks for developer adoptiondataset release
Deal Breakers
18+ years of experience in product management or data operations specifically within the AI/ML sector, 8+ years direct people management experience, Deep understanding of LLM/VLM architectures and alignment methods like RLHF and RLAIF, Bachelor's or Master's degree in Computer Science, AI/ML, Data Science, or related technical field (or equivalent experience)
Get matched to jobs like this
Luna finds roles that fit your skills and career goals — no endless scrolling required.
Create a Free Profile