Key Responsibilities
- Establish the canonical production architecture, repository, data model, and release path.
- Freeze and harden the three highest-value end-to-end client workflows.
- Close gaps in tenant isolation, integrations, testing, recovery, security, and capacity.
- Consolidate or retire duplicate queues, surfaces, scripts, and pipelines.
- Own production releases, canaries, rollbacks, migrations, incident response, and disaster recovery.
- Enforce provider quotas, cost budgets, idempotency, backpressure, and operational telemetry.
- Direct AI coding agents through bounded tickets and rigorously review their output.
- Translate product and content requirements into executable acceptance contracts.
- Produce an evidence-backed assessment of 1,000-client capacity and unit economics.
- Create a technical roadmap that clearly separates platform hardening from new feature development.
Required Qualifications
- Eight or more years building production software, including senior, staff, or principal-level responsibility.
- Personally shipped and operated a multi-tenant B2B SaaS product.
- Strong TypeScript, Next.js, Python, FastAPI, Postgres, SQL, and row-level security experience.
- Proven judgment in multi-tenant architecture, tenant isolation, secrets management, least privilege, security, and auditability.
- Deep experience with durable queues, leases, retries, idempotency, dead-letter handling, backpressure, and ambiguous-write recovery.
- Experience managing external-provider quotas, throttling, failures, and cost controls.
- Disciplined ownership of database migrations, production observability, incident response, rollback, and disaster recovery.
- Hands-on experience with load, soak, integration, and end-to-end testing.
- Effective use of AI coding agents without lowering engineering or review standards.
- A demonstrated willingness to simplify inherited systems and remove unnecessary duplication.
- This person must have personally operated production systems, not merely built features that someone else operated.