Snorkel AI builds specialized training data, benchmarks, and evaluation environments that help frontier models and agents perform in high-stakes domains. Backed by Greylock, Lightspeed and GV.
About the role
Frontier AI data is expensive to make and hard to measure. Every task we deliver is tested against the strongest models, often through many long-running agent rollouts. Your job is to make that process faster, cheaper, and more rigorous with ML and AI
What they're looking for
- Efficient agentic evals. Cut the cost of long-horizon agent evaluation with adaptive sampling, statistically grounded early stopping, model cascades, caching, and cheap-first gating
- AI model routing. Route every eval and judge call to the cheapest model that clears the quality bar, with fallback, monitoring, and cost attribution
- Fine-tuned small models. Fine-tune and serve open-weight models (LoRA and other parameter-efficient methods) where they match frontier quality, and know when they don't
- Predictive difficulty. Build models that estimate how hard a task is for frontier systems before running a single rollout
- Measurement for AI data. Build golden datasets, quantify the accuracy and calibration of LLM-as-judge systems, and make quality reproducible across projects
- Research to production. Turn research prototypes into reusable, configurable components that forward deployed engineers and researchers use on every project
More about this role
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!
In September 2026 we raised a $350 million Series E at a $3.5 billion valuation , and we are scaling our engineering and research teams to meet demand.
Frontier AI data is expensive to make and hard to measure. Every task we deliver is tested against the strongest models, often through many long-running agent rollouts. Your job is to make that process faster, cheaper, and more rigorous with ML and AI
You will be one of the early members of ML & Research Engineering at Snorkel. You will study how frontier-grade data is generated and evaluated, form hypotheses, validate them against real production data, and ship the winners at scale. You will shape the discipline's direction, its...
Browse similar: AI jobs · AI startup jobs · Startup jobs · San Francisco Bay Area