AI Engineer

About the Role

You'll own the patient-facing AI conversation loop end-to-end at Doctours, diagnosing and fixing failures when our AI agent stumbles with patients. This role bridges infrastructure and empathy—you'll build robust agent systems while staying connected to real patient outcomes. You'll work across tool integrations, evals, retrieval, and clinic-side automation, shipping code that directly improves care.

What You'll Do

• Diagnose why AI conversations with patients fail and ship code fixes to prevent recurrence.
• Build and maintain agent infrastructure including tool integrations, evaluation frameworks, and retrieval systems.
• Automate recurring manual workflows on the clinic side, eliminating toil through code.
• Handle patient-facing situations with empathy when the AI doesn't, turning those moments into product insights.
• Own systems end-to-end with minimal oversight, making architectural and implementation decisions independently.

What We're Looking For

• Demonstrated ability to build and ship LLM or agent systems end-to-end (professional experience, personal projects, research, or open-source all count equally).
• Experience with agent infrastructure: evals, RAG/retrieval, and tool integrations.
• Strong CS and math fundamentals; elite academic signals like top-tier university degrees, research, or math competition success are a strong plus.
• Automation instinct—you spot recurring manual workflows and proactively eliminate them through code.
• Patient-facing warmth and empathy; you can handle difficult conversations with care and extract product lessons from them.
• Startup autonomy: comfortable owning systems with minimal oversight and making decisions independently.
• Existing US work authorization (no visa sponsorship available).

Location

Full-time, in-person at our San Francisco headquarters, five days per week.