Research Scientist - Human-AI Systems

Overview

Snorkel AI is hiring a Research Scientist to help advance how high-quality data and environments for AI agents are created. This role focuses on building and improving pipelines that combine real-world data, automated generation, and human expert input to support frontier AI systems. You will work with domain experts, academic partners, customers, and internal product and engineering teams to scale these workflows and improve data diversity, data quality, and model performance.

This position is ideal for someone who wants to turn research into reusable, scalable systems for data-centric AI.

What You'll Do

• Design, implement, and improve reusable pipelines that combine AI capabilities with expert judgment for data and agentic environment creation.
• Run rigorous experiments to validate proof-of-concept ideas and measure their effect on data quality, pipeline efficiency, and model performance.
• Partner with engineering, data operations, domain experts, and customers to turn prototypes into reliable production workflows.
• Work directly with domain experts to design and test workflows that help them author, review, and refine data and environments.
• Collaborate cross-functionally to surface research findings that inform product and roadmap decisions.
• Stay current with research in data and agentic environment creation and bring relevant best practices into internal workflows.
• Represent Snorkel's research externally through publications, blog posts, conference talks, and customer engagements.

Preferred Qualifications

• Research background in AI, machine learning, NLP, LLMs, or a related field, with experience developing and evaluating new methods.
• Experience building environments for AI agents in automated research, computer use, coding, or professional-domain workflows.
• Experience with synthetic data generation, human-in-the-loop workflows, reinforcement learning, agent environments, or model evaluation.
• Strong experimental design skills, including forming hypotheses and running ablation studies.
• Software engineering practices such as clean coding, modular design, and version control.
• Ability to translate domain expert knowledge into concrete tasks, evaluation criteria, and repeatable workflows.
• Comfort working through rapid iteration, ambiguous research questions, and the path from experimentation to production.
• Ph.D. in machine learning, NLP, or a related field preferred; equivalent industry or research lab experience is also considered.