Member of Technical Staff, Agents

• Build agents that complete delegated work end to end with minimal user oversight.
• Define trustworthy measures of agent quality and evaluate performance across representative tasks.
• Develop new agent capabilities for tools people already use, and verify they are ready before launch.
• Investigate production failures and deliver durable, measurable fixes.
• Build systems to detect known failure patterns and less predictable issues at scale.
• Improve agent recovery when tools or websites fail.
• Create repeatable evaluation practices so engineers can assess whether agent changes improve outcomes.
• At least 3 years in data science or machine learning, building or delivering evaluation systems, metrics, or product quality measurement.
• Strong production proficiency in Python and SQL, including analysis of large datasets and automated pipelines.
• Experience designing evaluation frameworks with measurable criteria, representative datasets, and appropriate metrics.
• Experience building or operating agents for real users and improving their reliability.
• Familiarity with statistics, experiment design, and realistic end-to-end simulations.
• Experience owning systems through production launch and measuring their impact.
• Experience with post-training, fine-tuning, or reinforcement learning is relevant.
• Working knowledge of Django, React, and TypeScript is useful.
• Typically suited to candidates with fewer than 5 years of engineering experience and a track record of progression or startup ownership.
• On-site in Palo Alto, California, with four days per week in the office.

None flagged.