TypeScript Engineer for SaaS Reliability and Regression Prevention

We have an established multi-tenant B2B SaaS application built with React, TypeScript, Node.js, Express, PostgreSQL, Drizzle, Stripe, SendGrid, Vitest, and Playwright. The application is live and actively used. It was developed rapidly with significant assistance from Claude Code and Codex. The codebase has extensive documentation, CI checks, critical-path protections, and more than 8,000 automated tests. Our current challenge is regression control. Individual defects are being corrected, but some changes have unintentionally affected adjacent workflows. We need someone who can help us improve the reliability of the existing application without rewriting it or slowing development to a halt. The work would include: Investigating reported defects across the complete workflow Identifying root causes rather than patching visible symptoms Reviewing whether previous corrections fully addressed the problem Adding behavioral regression tests that exercise real functionality Strengthening end-to-end coverage for critical customer workflows Reviewing sensitive changes before they are merged Identifying duplicated or inconsistent business logic Improving release confidence and production-verification procedures The most important areas include: Invoicing and recurring billing Stripe payments, subscriptions, and webhooks Public signup, quoting, and customer conversion Scheduling, routing, and field-service workflows Email and SMS communications Database migrations and schema consistency Authentication and tenant-data isolation We are not looking for: A complete rewrite A new architecture proposed before understanding the existing system General feature development Cosmetic cleanup Someone who only reruns an AI prompt until the immediate error disappears Tests that inspect source-code text without verifying actual behavior We need someone comfortable inheriting a large, unfamiliar TypeScript application, reproducing defects, tracing data through the frontend, API, database, background jobs, and external integrations, and making narrow corrections supported by meaningful tests. We would begin with a paid five-hour trial on one contained regression. The goal of the trial is to evaluate the developer’s investigation process, technical judgment, testing quality, and communication. The trial would require: Review and reproduce one recent regression Trace the complete affected workflow Identify the underlying cause Implement or recommend the smallest safe correction Add behavioral regression coverage Document adjacent risks and anything that could not be verified Submit the work through a pull request The trial will not include production access, deployment authority, or permission to make unrelated changes. If the work is strong, this could continue as a part-time engagement focused on stabilization and reviewing high-risk changes.