About the Role

You'll own the agentic infrastructure at Orpex—the systems that let agents execute reliably, remember context over months, and know when to pull in a human. While everyone is building agents, almost nobody builds agents that actually run in production at scale. You'll ship to production in your first week and decide how agents plan, remember, fail, and recover.

What You'll Do

• Build and maintain the core infrastructure that enables agents to execute tasks reliably, even after dozens of iterations.
• Design and implement memory systems that retain months of client context from messy operational data.
• Develop eval harnesses that teams actually trust to decide what ships to production.
• Own the full loop: build it, measure it, break it, fix it, and make it learn.
• Ship agent systems for real customers and iterate based on production feedback.
• Handle whatever needs doing—infrastructure, orchestration, customer collaboration, and hiring.

What We're Looking For

• Shipped LLM agents that ran unattended in production for real users; you understand that the model is the easy part.
• Built eval harnesses that people actually used to make shipping decisions.
• Experience with retrieval and memory systems that work on real, messy operational data rather than demo datasets.
• Strong Python proficiency and hands-on experience with agent orchestration and infrastructure.
• Track record of intensity and execution—you force things into existence; bonus if you've built agent tooling in the open or founded something.

Compensation & Benefits

• $150–250K USD base + meaningful early equity.
• Relocation and visa sponsorship covered.

Location

San Francisco, in person.