Human Data Strategist (AI)

Location: On-site in New York, NY or San Francisco, CA, United States

Company

hirify.global is a research lab building open models designed to make intelligence accessible, customizable, and usable by everyone.

What you will do

• Translate research plans and critical model capabilities into executable data acquisition strategies.
• Assess and coordinate data vendors, including capacity forecasting, task instructions, acceptance criteria, quality, delivery, and cost.
• Design contributor workflows and frameworks for task and dataset quality.
• Build dashboards tracking throughput, turnaround time, acceptance rates, rework, and cost per accepted task.
• Use analytics to investigate bottlenecks and guide campaign improvements.
• Work with researchers to incorporate feedback and measure how acquired data improves model capabilities.

Requirements

• 4+ years of experience delivering complex program portfolios and prioritizing across changing research needs.
• Understanding of LLM training, evaluation, and prompt design.
• Proficiency in SQL for querying datasets and building dashboards.
• Proficiency in Python for data analysis, API integrations, and workflow automation.
• Clear written communication, strong stakeholder judgment, and willingness to inspect individual tasks and datasets in detail.
• High agency and ability to drive impact in fast-paced, dynamic, and ambiguous environments.

Nice to have

• Experience with annotation, expert data collection, human feedback, or AI evaluation.

Culture & Benefits

• Top-tier compensation with salary and equity designed to recognize and retain talent.
• Stock options for all employees.
• Comprehensive medical, dental, vision, and life insurance with an annual wellness allowance.
• Daily lunch and dinner provided in the office.
• 22 weeks of paid parental leave for all new birthing and non-birthing parents.
• Unlimited paid time off in the U.S., regular team events, and visa sponsorship with support for long-term immigration pathways where applicable.