Role Summary

You will own the strategy, team, and execution for frontier-grade data that powers cutting-edge AI systems. You'll build the organization, keep us on the bleeding edge of data methods, and ship proactive, off-the-shelf data packs and tooling that measurably improve model quality and time-to-value.

What You'll Do

• Engage with leading LLM labs to advance LLMs across STEM domain
• Engage with a team of leading STEM experts in US and across the globe to generate AGI advancing data in STEM
• Understand and define what constitutes good STEM data (data diversity, model breaking prompts, pass @K distribution)
• Define and understand data quality rubric
• Drive data generation operations through PMO reporting into this role
• Responsible to build a team of Strategic project leads (SPL) and oversee various client data generation workstreams being run by the SPLs
• Set headcount plans, budget, vendor strategy, and capacity models that scale.
• Understand SFT/RLHF/RLAIF/Evals methodology
  Define and continuously refine quality definitions and measurement (rubrics, gold sets, adjudication, inter-annotator agreement, automated checks, eval harnesses).
• Produce offering collateral and internal research briefs that convert into real customer value.
• Ship proactive data packs
• Own the roadmap and production of off-the-shelf data packs (by domain, modality, and task); ensure packaging, documentation, licenses, and release notes are crisp.
• Drive cross-company learning: postmortems, playbooks, and pattern libraries so wins compound.
• Tools, generation, and proof of value
• Stand up proactive data generation (human + synthetic), QC tooling, dashboards, and auto-checks integrated into CI for data.
• Lead fast PoV cycles with customers: sample packs, eval notebooks, and "time-to-first-signal" demos.
• Close the loop with customers
• Systematically implement feedback from Frontier Data Managers and Sales; translate signals into roadmap changes, SLAs, and new pack definitions.

Minimum Qualifications

• STEM degree (CS, Math, Stats, EE, or related). Advanced degree preferred.
• 10+ years in data/ML/analytics or data product roles; 4+ years leading managers/leads in high-growth environments.
• Deep command of the data lifecycle for AI systems: sourcing, labeling, synthesis, QA, evals, and deployment feedback loops.
• Hands-on fluency with Python/SQL and modern data/ML stacks (cloud object stores, distributed compute, labeling/QC systems, experiment/eval frameworks).
• Track record of turning research into shippable data products and measurable quality lift.
  
  

Nice to Have

• Experience with RLHF/RLAIF pipelines, multimodal data, agents/tool use, and safety evaluations.
• Built or operated human-in-the-loop programs at scale (onshore/offshore) with rigorous QA.
• Familiarity with data licensing, IP, and safety/privacy constraints for AI training.

Core Competencies

• Org builder with crisp operating cadence and high hiring bar.
• Product sense for data: knows when a dataset is "sellable" and how it will move eval metrics.
• Metric-driven and experimental: designs evals that survive contact with reality.
• Clear communicator who earns trust with execs and ICs.
  
  

Success Metrics (12 Months)

• Data Packs: 6-10 new packs launched; ≥70% reused across ≥3 customers each.
• Quality Lift: ≥X% improvement on target evals (task-specific) vs. baseline; ≤Y% variance.
• Talent: >90% retention for top performers; time-to-productivity ≤60 days.
• Voice of Customer: Closed-loop feedback implemented with measurable NPS/CSAT lift.

How We Work

• High-ownership, low-drama execution.
• Ship small, validate fast, scale what works.
• Ruthless clarity on quality definitions and measurement.

Compensation Range for this role - USD 350,000 to USD 410,000