SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Patronus AI is a frontier AI lab developing simulation research and infrastructure to accelerate progress toward human-aligned AGI. The company is behind influential AI evaluation research including FinanceBench, Lynx, SimpleSafetyTests, CopyrightCatcher, and Humanity's Last Exam. Founded by former researchers from Meta AI, Amazon AGI, and Google, Patronus serves foundation model labs and Fortune 500 enterprises like Adobe, backed by investors including Lightspeed Venture Partners and Stanford University.
As Technical Program Manager, you will own the end-to-end delivery pipeline that transforms customer work orders into shipped RL environments and simulations. You manage the complete lifecycle from signed contract through SME sourcing, environment engineering, task generation, QA, and final delivery to customers. This role requires you to maintain clarity on ownership, timelines, and readiness across engineering, QA, and subject-matter-expert teams while navigating constant scope changes—app lists expanding mid-project, difficulty definitions being renegotiated weeks before delivery.
This is not a coordination-only role. You stay deeply involved in the details: reading QA feedback, testing environments, sanity-checking task quality, and pitching in directly where needed. You identify automation opportunities and build them yourself. You define what "ready to ship" means and enforce it—QA tickets must be green and tasks must pass in the customer's harness before delivery, not after.
Key responsibilities include: owning delivery programs end-to-end with clear tracking of work orders and deliverables; managing scope changes and updating plans, pricing, and commitments without losing the thread; running handoffs between GTM, SME sourcing, engineering, task generation, and QA to eliminate stranded work; working directly with customers to turn asks into plans with clear outcomes and owners; ensuring teams time-bound their work and make estimates explicit; building tools and scripts that scale your function; and staying hands-on with agent testing, QA review, and small-tool development.
Your work directly impacts frontier labs' ability to stress-test and improve the next generation of AI agents, advancing progress toward safe, human-aligned general intelligence.
