• 5+ years building backend and infrastructure systems, with strong Python
• You have owned production systems end to end: deploys, databases, observability, and the pager. You have been the person who got paged, and you fixed the class of problem rather than the incident
• You like adversarial problems. Systems where something on the other side is actively working against you
• You take correctness seriously because the stakes are real. These systems spend customers' money, so a silent failure is a wrong order, not a red build
• You can move without a finished spec. We are small, and the roadmap changes when the market does
• Built AI-first products or LLM-based agents in production, not just in a prototype
• Working knowledge of classic ML and deep learning, enough to reason about retrieval and ranking rather than only calling an API
• Background in e-commerce, payments, or fraud and risk infrastructure