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 a Forward Deployed Engineer focused on Computer Use Agents to partner with leading AI labs and enterprises on their most critical agentic-AI initiatives.
In this role, you will lead the technical execution of complex customer engagements involving agents that operate computers, browsers, and software environments to complete realistic, multi-step tasks. You will translate ambiguous product and model challenges into robust task environments, datasets, evaluators, and delivery plans that improve agent reliability and downstream performance.
You will work across the full delivery lifecycle—from technical discovery and solution design through implementation, evaluation, and production delivery. You will also identify patterns across engagements and turn successful approaches into reusable capabilities, technical standards, and product improvements.
Main Responsibilities
Computer Use Agents, Data, and Evaluation
- Design and build task environments, datasets, and evaluation workflows for computer-using agents operating across browsers, desktop applications, terminals, and other software interfaces
- Translate customer goals, agent failure modes, and real-world workflows into representative, multi-step tasks with clear success criteria
- Develop data-generation, validation, and quality-assurance pipelines for multimodal and agentic training and evaluation data
- Build automated evaluators, checks, and measurement frameworks to assess task completion, correctness, robustness, efficiency, and adherence to requirements
- Diagnose agent failures across planning, tool use, perception, state management, and interaction with user interfaces; turn findings into improved tasks, data, and evaluations
- Design and run experiments to measure how data, task design, and evaluation changes affect downstream agent performance
- Deliver reusable, production-grade task suites, datasets, and evaluation assets that help customers train, benchmark, and improve computer-use agents