About this role
About the RoleBuild and improve autonomous AI agents that handle real executive workflows, where reliability matters. You will work across agent development and evaluation, identifying production failures and proving that fixes and new capabilities perform well before they reach users.
What You'll DoBuild agents that complete delegated work end to end, without requiring users to check every step.
Develop reliable ways to measure how often agent tasks are completed correctly.
Design new agent capabilities and validate their readiness before launch.
Investigate production failures and implement fixes that address underlying system issues.
Build failure detection that catches known issues and emerging patterns at scale.
Improve agent recovery when tools or websites fail.
Make agent changes measurable and repeatable so engineers can assess their impact.
What We're Looking ForExperience building or improving production AI agents or other applied machine learning systems.
Background in evaluation, harness engineering, or measuring the reliability of AI systems.
Strong Python skills and familiarity with Django, React, and TypeScript.
Comfort investigating ambiguous problems, working across system components, and connecting technical improvements to user outcomes.
Familiarity with AI evaluation tools such as Braintrust or Raindrop is relevant.
Compensation & BenefitsSalary range: $200,000 to $300,000 USD annually. Visa sponsorship is not available.
LocationOn-site in Palo Alto, California, United States, with four days a week in the office.