Senior Research Scientist, Frontier AI Data & Benchmarks

Snorkel AI; New York (NY); On-site; USD 150,000 - 210,000; Full time

Job summary

Snorkel AI seeks a Staff or Senior Research Scientist to lead dataset development for frontier benchmarks. The role is highly visible and customer-facing, bridging research, strategy, and GTM.

You will design datasets aligned with frontier model performance and collaborate with academic partners, then scale production with delivery, product, and marketing teams. The position reports to the Head of Research and suits someone energized by cross-functional work and startup dynamics.

Qualifications

• Strong research background in AI/ML evaluation, NLP, or related fields.
• Exceptional communication skills for technical and non-technical audiences.
• Comfort in fast-moving cross-functional environments with ambiguous problem spaces.
• Genuine interest in GTM strategy, startup dynamics, and commercial AI data services.

Responsibilities

• Design state of the art datasets for frontier model training and evaluation.
• Translate benchmark insights into clear narratives for ROI in customer-facing materials.
• Collaborate with data operations, product, engineering, and strategy to inform company roadmap.
• Stay at the frontier of LLM evaluation research and integrate best practices into workflows.
• Represent Snorkel's research externally through publications, talks, and engagements.

Skills

Research evaluation

Scientific communication

Cross-functional collaboration

Ambiguity management

GTM interest

Education

PhD in ML/NLP or related

Job description

Snorkel AI seeks a Staff or Senior Research Scientist to lead dataset development for frontier benchmarks. The role is highly visible and customer-facing, bridging research, strategy, and GTM.

You will design datasets aligned with frontier model performance and collaborate with academic partners, then scale production with delivery, product, and marketing teams. The position reports to the Head of Research and suits someone energized by cross-functional work and startup dynamics.