Keystone Bookkeepers
AI-First Product & Platform Engineering Location Remote — Latin America Hours 11:00 AM – 7:00 PM ET Employment Type Full-Time About The Role Keystone Bookkeepers is a bookkeeping and accounting firm specializing in short-term rental operators and property managers. We are not a software company that happens to do accounting — we are an accounting firm building the software our own operation runs on, which means the things you build get used by real bookkeepers and real clients within days, and you get to see whether they worked. We are hiring a senior, AI-first software engineer to be the accountable technical owner of our internal platform and client-facing tooling. You will design, build, and operate production applications, using tools like Claude Code and other agentic coding platforms as a core part of your daily workflow — not as an occasional assistant. This is not a greenfield role and it is not a feature-ticket role. You are inheriting live systems that handle client financial data, and you will be accountable for them from design through implementation, deployment, monitoring, and maintenance. When something breaks during the business day, you are the person who finds it and fixes it. You will work alongside an existing engineer based in the Philippines and an external development partner. You are the senior technical voice in the room and the escalation point during North American business hours. What You'll Own Production reliability during North American business hours — you are the first responder, not the escalation of last resort End-to-end delivery of features across our internal operations platform and client-facing portal Integration health across QuickBooks Online, Microsoft Graph, Bill.com, property management platforms, and CRM systems The security posture of everything you ship, with explicit acceptance criteria rather than a post-hoc review Operational visibility — if a client notices a problem before we do, that's a gap you close Key Responsibilities Design, develop, test, deploy, and maintain scalable web applications and backend services Build AI-enabled features and integrate LLM APIs where they add real value — document extraction, classification, review workflows, and agent-driven automation Use AI development tools (including Claude Code or similar) as part of your daily engineering workflow to increase both productivity and software quality Build and maintain REST APIs, serverless functions, background workers, scheduled jobs, and webhook ingestion pipelines Own database work end to end: schema design, migrations with a tested rollback path, and query performance Triage and resolve production incidents across the application, database, and third-party integration layers Instrument the system so failures surface on their own — structured logging, alerting on authentication and integration failures, and monitoring of queues and background jobs Write and maintain runbooks so recurring problems become procedures other people can execute Extend and modify existing codebases you did not write, without unnecessary rewrites Write clean, maintainable, documented code and participate in code review Translate technical risk and tradeoffs into business terms for firm leadership and operations staff Required Qualifications Engineering Bachelor's degree in Computer Science, Software Engineering, or equivalent practical experience Senior-level experience across the full stack, with a demonstrable track record of shipping and then operating what you shipped Strong proficiency in JavaScript/TypeScript and React, building modern frontend applications Experience developing Node.js / serverless backends. Familiarity with React, Vite, Tailwind CSS, Vercel Serverless Functions, and Supabase/PostgreSQL is highly desirable Solid experience with SQL/PostgreSQL — data modeling, indexing, query performance, and migration authoring Experience with REST APIs, webhooks, and third-party integrations, including handling provider API changes andfailure modes Experience with Git, testing, CI/CD, and collaborative development workflows Strong analytical, problem-solving, and written communication skills AI-First Development Genuine AI-first development experience . You regularly build software using Claude Code or comparable AI coding agents and can point to real projects you have shipped with that workflow. You understand effective prompting, when to intervene, and how to critically review AI-generated code — particularly in code paths that touch money or sensitive data. Production Ownership You have carried responsibility for a production system — taken the alert, diagnosed the incident, rolled back or shipped the fix, and closed the gap afterward You have inherited and extended a codebase written by someone else, and can describe how you got up to speed on it You have set up alerting and logging, not just consumed someone else's dashboards Comfortable being the accountable owner from design through secure implementation, deployment, monitoring, and long-term maintenance Security & Sensitive Data Our platform handles client financial records, tax documents, trust account data, and third-party API credentials . Security is a product requirement here, not a late-stage checklist. Experience building or maintaining software that handles sensitive financial or otherwise regulated data Working knowledge of multi-tenant data isolation — PostgreSQL row-level security, scoped request context, and authorization enforced at the server boundary rather than in the UI OAuth 2.0 integration experience including full token lifecycle: refresh workers, expiry and revocation handling, and encrypted credential storage Secure document handling — access-controlled storage, access auditing, and avoiding leaked signed URLs Secure SDLC fundamentals: input validation, IDOR prevention, secrets hygiene, and dependency vulnerability triage Infrastructure Competence You do not need to be an infrastructure specialist. You do need to be able to keep the lights on without help. Docker in production — multi-stage builds, deploys, and a rollback strategy Hands-on experience operating at least one major cloud platform (Azure, GCP, or AWS) You have personally restored a PostgreSQL database from a backup Able to read and modify a reverse proxy configuration and manage TLS certificate renewal Availability Consistent availability 11:00 AM – 7:00 PM ET, Monday through Friday The first two hours of your day overlap with our Philippines-based engineer for daily handoff; the back half of your day is North American coverage, which is when client-facing issues surface. Both halves matter, and the schedule is not flexible on this point Preferred Qualifications Domain experience in accounting, bookkeeping, fintech, or property management software , with the ability to reason about financial data Experience integrating with QuickBooks Online or a comparable accounting/ERP platform Applications built on Anthropic Claude API or OpenAI API Deep familiarity with Claude Code specifically Proficiency in Python for automation, data synchronization, and background processing Schedule d jobs, cron workflows, GitHub Actions, or other background processing systems Exposure to threat modeling, access logging, or data retention policy work Experience in a SOC 2 or similarly compliance-minded environment Experience working effectively in a distributed remote team across multiple time zones Your First 90 Days We would rather be specific than aspirational. Here is what success looks like. Days 1–30 — Absorb the escalation load Take ownership of the afternoon issue queue; problems route to you first Get our existing bug backlog into a real tracker and triaged Write runbooks for the top recurring failures Days 31–60 — Make failures visible before clients see them Alerting in place for integration failures, invoice and billing job failures, and background job errors Backup restore verified rather than assumed Post-deploy verification in place for releases Days 61–90 — Own real scope Take full ownership of at least one significant workstream Ship a meaningful feature end to end, with security acceptance criteria defined up front What We Offer Competitive salary and benefits Significant ownership and a direct line to the founder — no layers between you and the decision A collaborative, remote-first environment across three regions The opportunity to build modern AI-powered software with current tooling, and to decide how it gets built Meaningful technical problems: multi-tenant financial data, accounting integrations, and AI-assisted workflows that replace genuine manual effort Ongoing professional development and continuous learning How To Apply Submit your resume along with a brief note covering: Projects you have built using Claude Code or similar AI coding tools, and how AI has changed your workflow A production incident you owned — what broke, how you found it, how you fixed it, and what you changed so it would not happen again A codebase you inherited from someone else and what you did with it Links to your GitHub, portfolio, or examples of work, if available We read these. A short, specific note will get further than a long, general one
Keystone Bookkeepers
