Director of AI Research & Data Operations

Cybersecurity & Trust and Safety | RLGym

We are looking for a Director of AI Research & Data Operations to lead and scale RLGym's research and data generation efforts, delivering high-quality training and evaluation data to frontier AI models at scale.

This role combines research leadership, data science, and large-scale operational management. You will build and lead multidisciplinary teams of researchers, data scientists, domain experts, and operations specialists, transforming complex research challenges into scalable, high-quality data production systems.

The ideal candidate brings experience managing large research and data operations, a strong understanding of modern AI and LLM development, and the ability to deliver high-quality datasets that advance frontier model capabilities, with a particular focus on cybersecurity and trust & safety.

What you'll be doing

• Lead and scale large AI research and data operations, managing multidisciplinary teams of researchers, data scientists, domain experts, and operational leaders.
• Own the end-to-end generation and delivery of high-quality training and evaluation datasets for frontier AI models, ensuring scalability, consistency, and research value.
• Define the research strategy, operational roadmap, and data generation priorities across cybersecurity, trust & safety, and emerging AI capabilities.
• Build and optimize large-scale data production pipelines, combining human expertise, automated workflows, synthetic data generation, and AI-assisted research.
• Establish rigorous data quality frameworks, including validation methodologies, quality assurance processes, benchmarking, and continuous improvement.
• Lead research initiatives focused on generating realistic, challenging, and diverse datasets for LLMs, GenAI applications, and agentic AI systems.
• Work closely with research, product, engineering, and data science teams to translate frontier model requirements into actionable research programs and scalable data deliverables.
• Develop scalable systems for expert-driven data generation, evaluation, annotation, and human feedback, including workflows supporting supervised fine-tuning and reinforcement learning.
• Define and track operational and research KPIs, balancing data quality, throughput, cost efficiency, and delivery timelines.
• Build partnerships with external researchers, domain experts, and research organizations to expand data generation capabilities and domain expertise.
• Drive innovation in data generation methodologies, evaluation benchmarks, and research infrastructure to continuously improve the quality and impact of RLGym's data.

What we need to see

• Proven experience leading and scaling large research, data science, AI/ML, or data operations teams.
• Strong experience managing complex, high-volume data generation or research operations, with measurable quality and delivery outcomes.
• Deep understanding of AI/ML development, LLMs, GenAI, and the role of training and evaluation data in improving model performance.
• Experience designing and implementing data quality frameworks, evaluation methodologies, and scalable research workflows.
• Ability to translate complex research objectives into structured execution plans and deliver results at significant scale.
• Strong analytical and data-driven decision-making skills, with experience defining operational metrics and optimizing research processes.
• Experience managing multidisciplinary teams and collaborating with research scientists, engineers, and domain experts.
• Strong strategic leadership, operational execution, and cross-functional communication skills.
• Familiarity with cybersecurity, AI security, or trust & safety challenges and their relevance to frontier AI model development.
• 8+ years of relevant experience in AI/ML, data science, research leadership, large-scale data operations, or related technical fields.

Ways to stand out

• Experience delivering large-scale training or evaluation datasets to leading frontier AI labs.
• Background in human feedback, RLHF, reinforcement learning environments, synthetic data generation, or expert-driven annotation programs.
• Experience building data generation platforms, automated evaluation pipelines, or large-scale AI research infrastructure.
• Familiarity with cybersecurity and trust & safety datasets, adversarial evaluations, or AI red-teaming methodologies.
• Experience developing data quality measurement systems, including automated grading, expert calibration, and statistical quality control.