Backend Developer for Convex, TypeScript & Data Modeling

We're looking for a backend specialist to join our team on an ongoing basis. This role is for someone who thinks in data models, reactivity, and clean function design, and who can build a backend that stays maintainable as it grows. We are a service-as-a-software company. Our engineering team works closely with and for our ops teams to build the best service-as-a-software product possible. You'll be modeling real operational workflows and building the backend that powers them, in direct collaboration with the people who use it. BEFORE YOU APPLY: We work exclusively with independent developers. Any agency that applies will be automatically rejected. We will ask to see real examples of backend work, ideally Convex projects, or comparable reactive/serverless backends with thoughtful data modeling. Please have those ready before reaching out. STACK YOU MUST KNOW: TypeScript (strong; our backend is fully typed end-to-end) Convex (this is the core of the role; deep, hands-on experience required) WHAT WE'RE. LOOKING FOR: A backend developer with real depth in Convex and a strong instinct for data modeling. You'll own the data layer the rest of the product sits on. CORE CONVEX SKILLS: - Fluent writing queries, mutations, and actions, and knowing when to reach for each - Designing schemas that are clean, normalized where it matters, and built to evolve - Modeling relationships and access patterns thoughtfully (one-to-many, many-to-many, tenancy) - Reasoning about indexes and query performance, avoiding full table scans and paginating correctly - Mastery of Convex's reactive model, understanding how queries invalidate and re-run, and designing for it - Server-side validation with validators, and making invalid states unrepresentable via the type system - Auth & identity flows, multi-tenancy, and access control done correctly and securely - Scheduling, cron jobs, HTTP actions, file storage, and webhooks where needed - Structuring backend functions across a growing codebase for maintainability, with composable internal functions, clear module boundaries, and no one-off hacks ENGINEERING DISCIPLINE: - Strong TypeScript, typed end-to-end with no any escape hatches, and types that document intent - Writing tests as you build, especially around auth, tenancy, and provisioning - Thinking about data integrity, migrations, and backwards compatibility - Performance- and cost-aware (bandwidth, function calls, query efficiency) BONUS / NICE-TO-HAVE: - Experience with multi-tenant / multi-module architectures - Auth providers and identity federation (e.g. WorkOS-style JWT validation) - Building backends for AI-native products, including agent state, tool execution, streaming, and background jobs - Comfort bridging into the frontend (Next.js) enough to design APIs the UI actually wants to consume HOW WE WORK: - Direction, principles, and implementation conventions are heavily shaped by our pre-built set of gen-AI plugins, skills, and commands. They're there to support you, not box you in. - Much of how things get done is already encoded in our internal tooling, so you spend less time on boilerplate and setup and more on building - The conventions give you a strong, opinionated foundation to move fast and stay consistent with the rest of the team - Part of the job is leaning on these tools, then helping improve and extend them as you go AI-POWERED DEVELOPMENT: A MANDATORY REQUIEREMENT: - We are an AI-first team and expect every developer we work with to operate the same way. - Proficiency with generative AI coding tools is not a nice-to-have. It is a hard requirement. - Specifically, we require fluency with agentic coding tools like Claude Code and Codex, including: - Best practices around context management - Prompt engineering for development tasks - Using MCPs (Model Context Protocol servers) to extend and accelerate workflows - You should be using these tools daily to speed up development, assist with code review, scaffold features, and maintain quality. Developers who are not actively working this way will not be a good fit for this role. If this sounds like a fit, tell us a bit about yourself and drop links to relevant work.