Staff Engineer, AI Platform

## About the role We're building AI systems that actually work in production—not demos, not prototypes. Our platform processes hundreds of thousands of financial documents for enterprise clients, combining document intelligence with agentic workflows that make real decisions. We need a senior engineer who can operate across the entire stack, ship fast, and help lead a small team through high-stakes client work. This is not a role where you'll spend months in planning meetings. You'll be in the code, in client calls, and in production—often on the same day. ## What we're building Kodexa transforms how financial services organizations handle unstructured data. We're not bolting AI onto existing workflows—we're rebuilding document processing from the ground up with agents that can reason, extract, and act. Our clients include major banks and Fortune 500 companies processing everything from SEC filings to utility bills to freight documents. When our systems make a mistake, real money moves wrong. When they work, we eliminate thousands of hours of manual work. ## Responsibilities - Build across the stack. Our platform spans Python services, TypeScript/Vue user interfaces, and Java and Go backend services. You'll work wherever the problem is, not wherever your title says you should be. - Ship agentic AI to production. Design and implement AI workflows that combine LLMs, document extraction, and human-in-the-loop review. Make them reliable at scale. - Own production systems. You'll be on the hook for systems that process real client data. That means observability, performance, and the kind of paranoid reliability thinking that comes from experience. - Engage directly with clients. Join calls with engineering and operations teams at major financial institutions. Understand their problems firsthand. Translate messy real-world requirements into working systems. - Lead without a large team. We're small. You'll influence architecture, mentor other engineers, and help set technical direction—but you'll also write a lot of code yourself. ## Requirements - 10+ years shipping software, with real depth in at least two of: backend systems, data pipelines, frontend/UI. You've seen enough systems fail that you build differently now. - Production AI experience. You've put ML or LLM-based systems into production and dealt with the reality of model behavior at scale—not just training metrics. - Full-stack capability. You're genuinely comfortable in Python, TypeScript, and ideally Java and Go. You don't need to be an expert in all four, but you can't be afraid of any of them. - Client-facing presence. You can explain technical tradeoffs to non-technical stakeholders and hold your own in a room with enterprise clients. - Startup intensity tolerance. We're well-funded and have real revenue, but we move fast and priorities shift. You need to thrive in that, not just tolerate it. ## What we offer Competitive salary + significant options package (we'll talk specifics early in the