About Snorkel
Snorkel AI is the frontier AI data lab, helping teams build the data and environments behind high-performing frontier and agentic AI. We combine technology with research-driven AI data development to create datasets, benchmarks, evals, and custom solutions for real-world AI systems. Founded out of the Stanford AI Lab in 2019, Snorkel works with leading AI labs and enterprises to move from better data to better outcomes.
Excited to help us redefine how AI is built? Apply to be the newest Snorkeler!
About the Role
Snorkel AI is hiring Forward Deployed Researchers (Multiple levels) who will partner with frontier AI labs on their most challenging data problems. This is a high-impact, customer-facing role that combines deep technical and research credibility with the ability to work directly with customer research teams. You'll partner with customer research teams to design complex data and environments that improve frontier model performance, serving as a thought partner throughout the engagement.
You'll work at the critical intersection of research, technical strategy, and customer partnership. This includes scoping training data needs, designing RL environments and tasks, developing evaluation frameworks, probing model behavior and failure modes, and translating customer research objectives into actionable technical plans.
Forward Deployed Researchers own a technical domain. You are the person who knows where frontier models fail in it, what data moves those numbers, and what we should build next.
Main Responsibilities
- Own a technical domain end to end: the technical bar for what we build in it, and what we build next
- Partner with frontier AI research labs to design datasets and environments that improve model performance
- Lead technical conversations with customer researchers to understand model capabilities, failure modes, data requirements, and success criteria
- Probe model behavior through systematic evaluation to uncover weaknesses and identify high-impact data interventions
- Benchmark frontier models against our datasets to demonstrate value to customers, building or adapting evaluation harnesses as needed, and turn the results into pass-rate analysis, failure-mode breakdowns, and purchase recommendations
- Design evaluation frameworks, calibration processes, and quality rubrics that establish measurable project success metrics
- Develop technical specifications for data projects that balance research rigor with operational feasibility
- Hold the demand shape for your domain: which labs want what, at what volume, against what bar, and where the asks are converging
