Software Engineer

Halcyon Labs protects life’s largest transactions from fraud. Every year, more than $10B in payment fraud impacts U.S. real estate, and a single redirected wire can cost a family their home purchase or significantly impact a title company’s business. We operate at the intersection of identity, payments, and fraud prevention, serving the title, escrow, and lending markets. In 2025, Halcyon Labs protected 1.46 million real estate transactions. Our product continues to grow, our customers continue to expand with us, and our product engineering team plays a central role in building the closing experience that supports this growth. Why this role, why now We are building the Trust Engineering team as an agentic-first product engineering team. The operating model is straightforward: humans steer, agents execute, and humans verify. Engineers own architecture, product judgment, and the quality bar. AI agents support structured execution, including first-draft implementation, tests, documentation, and review preparation. The work starts with a clear specification, the agent builds against it, and the engineer reviews, verifies, and ships the final result. We are looking for an engineer who is already comfortable working this way, or is motivated to learn. This person should have a strong bias for action and be able to take an early-stage requirement, clarify it with product, and drive it to a reliable shipped feature through tight feedback loops. This is a hands-on product engineering role. Success is measured by customer impact, production quality, and shipped outcomes. How You'll Work Spec first. Every meaningful feature starts with a written spec: the problem, the desired outcome, scope and non-scope, edge cases, and acceptance criteria. You write clearly enough that an agent can implement against it and a teammate can review it. Steer the agents. You’ll use Claude Code and similar tools as part of your primary execution workflow. You guide the agent through implementation, generate and strengthen tests, and prepare clean PRs. You review AI-generated code critically, identify gaps, and own the result as if you wrote every line. Tight loops. You ship small PRs, behind flags, and iterate quickly. You prefer getting a real feature in front of a customer, learning from it, and improving it. You close the loop by monitoring production behavior, addressing issues, and moving the work forward. Compounding knowledge. You contribute to the team’s shared memory, including repo context files, prompt templates, decision records, and review checklists. The goal is for the team’s tools, agents, and workflows to improve over time because the way we build is documented and reusable. What You'll Own End-to-end features. You own your work from the database to the user interface, across the customer-facing web app, APIs, partner integrations, and workflows that support files from open to disbursement. You take a problem from rough idea through spec, build, review, and launch. Driving requirements to closure. You partner directly with the product, turn ambiguous asks into concrete specs, surface tradeoffs, and drive the work to completion. You do not wait for a fully defined ticket. You gather context, make informed decisions, and move the work forward. Quality you can stand behind. Money and identity flow through this product, so correctness is a core part of the role. You write tests, account for failure cases, and maintain a high quality bar on every PR, whether the first draft was written by you or by an agent. Production readiness. You ship, observe, and respond. You improve the codebase as you work and help raise the team’s standard for reliability, maintainability, and operational quality. What We're Looking For 3 to 5 years of professional software engineering experience shipping production web applications. A strong bias for action. You can break a large problem into shippable pieces, get something real in front of users quickly, and iterate. You are comfortable with ambiguity and can make informed decisions without waiting for perfect certainty. High ownership. You can take a requirement and drive it to closure without needing to be managed through every step. You think about the customer experience and the business outcome behind the work, not just the task itself. Fluency with AI-assisted development. You have experience building with tools such as Claude Code, Copilot, Cursor, or similar tools. You can review AI-generated code, identify issues, and own the final output. Real full-stack range. You are comfortable working in a modern front-end framework such as React with TypeScript, as well as in backend services. You treat both sides of the stack as part of the product experience. Strong written communication. You can write a clear spec, design note, or PR description that another person, or an agent, can act on. Written communication is central to how this team operates. Solid fundamentals. You have a strong understanding of HT