Sr Full stack developer

Job highlights

• Write and deploy serverless functions (Deno / Edge) for webhooks,scheduled jobs,and AI batch pipelines

Job Summary

Youll join a small, senior team building a multi-tenant operations platform for MSPs. The platform sits between the systems our partners already run their businesses on - PSA platforms like ConnectWise, HaloPSA, and Autotask - and layers on AI-driven analytics, customer health scoring, revenue intelligence, and a real-time ticket synchronization engine.

This is not a CRUD-app job. The core of the work is synchronization: keeping thousands of ticket pairs, notes, time entries, and attachments consistent across systems we don t control, with rate limits, flaky webhooks, and partners who expect zero data loss. On top of that data, you ll build the dashboards and analytics that make it useful.

Small team, real ownership, and a product where correctness is the feature. If you like problems that stay interesting after the first week, this is one.

Responsibilities

• Build and maintain the ticket sync engine - bidirectional ticket, note, time-entry, and attachment sync across ConnectWise, HaloPSA, Autotask, and ServiceNow, with webhook-first delivery, polling fallback, idempotency, and loop prevention.

• Design Postgres schemas and Row-Level Security policies that enforce strict per-tenant isolation inside a shared database.

• Ship React + TypeScript features - operational dashboards, admin configuration panels, and partner-facing analytics.

• Write and deploy serverless functions (Deno/Edge) for webhooks, scheduled jobs, and AI batch pipelines.

• Own data quality: deduplication, conflict resolution, drift detection, and self-healing reconciliation.

• Instrument everything - structured logs, job queues, dead-letter handling, and operational runbooks. You will be on call for the systems you build.

• Document what you ship. Every feature gets KB documentation. This product is sold on trust, and trust is written down.

  Must-haves

• React 18 + TypeScript 5 in real production apps.

• You can reason about data fetching, caching, and render performance without hand-waving.

• Postgres at depth. RLS policies you ve actually written, SECURITY DEFINER functions, partial and expression indexes, window functions. You ve debugged a recursive RLS policy and won.

• Supabase in production (or an equivalent Postgres + auth + storage stack), including auth flows, edge functions, and cron scheduling.

• Distributed-sync instincts. You can answer, without notes: two systems update the same record - how do you avoid duplicates and loops and a webhook and a poll race each other - what breaks

• API hardening. You ve integrated third-party APIs with pagination quirks, picklist validation, rate limits, and permission errors - and you read the docs closely enough to know that net_total isn t what it says it is.

• Production discipline. You validate every migration for regressions, write tests for sync edge cases rather than happy paths, and push back when a request will break something.

• Git fluency for managing a shared codebase. Branching and merge strategy, meaningful commits and readable history, rebasing without losing work, resolving non-trivial conflicts, and code review as a two-way conversation. You can bisect your way to the commit that introduced a regression, and you know how to revert safely in a live system at 2am.

• AI-assisted development as part of your normal workflow. You use tools like Claude Code, Cursor, or Copilot daily, and you can articulate where they speed you up and where they quietly cost you time.

• Clear written English and comfort working async with colleagues and partners in North America.

  Working with AI tools

We build AI features, and we use AI tools to build. We re not looking for enthusiasm - we re looking for judgment. Concretely, that means things like:

• Scaffolding an integration client from a PSA vendor s API docs, then hardening it yourself for pagination quirks, rate limits, and permission errors the model didn t know about.
• Generating a test matrix for sync edge cases - duplicate webhooks, out-of-order updates, partial failures - and pruning the cases that don t reflect real failure modes.
• Drafting a migration or an RLS policy with a model, then reading every line before it touches a shared database.
• Orienting in unfamiliar code - summarizing a module or a vendor SDK before you change it, rather than instead of understanding it.
• Turning a merged diff into a KB article.

Disclaimer: This job posting has been aggregated from external source. Role details, content, and availability are subject to change. Applicants are advised to confirm the latest information directly on the company website before applying.