What you'll own

• Temporal workflow definitions for the lien compliance lifecycle — state transitions, human-in-the-loop approval gates, activity retries, workflow versioning for in-flight migrations
• FastAPI service layer exposing shared business logic via HTTP — audit trail, fee validation, FMS sync endpoints
• FMS integration abstraction layer (SiteLink, storEDGE, Tenant Inc, SSM Cloud, Cubby) — partner-agnostic interface, raw data persistence, field-level variance detection
• Compliance reconciliation engine — validates FMS sync output before it enters any workflow; your code is the last gate before a legally invalid notice gets dispatched
• PostgreSQL schema design — append-only compliance event tables, audit trail permanence, sync run logging

Requirements

• 4+ years Python in production — not scripting, not data pipelines; service-layer backend code
• Temporal.io or equivalent durable workflow engine (Conductor, Cadence, Prefect with durability guarantees) — you must understand activity/workflow separation, determinism constraints, and versioning strategies for in-flight workflows
• FastAPI or equivalent async Python framework — you understand what async does to your database connection strategy, not just that it exists
• PostgreSQL — schema design, query optimization, migration discipline; you've written append-only audit schemas before or understand why you'd need to
• AWS — ECS, RDS, SSM Parameter Store, GitHub Actions CI/CD; you can read a Terraform file and reason about what it does

Highly Preferred

• Built integrations against external APIs with unreliable data — you've handled silent failures, blank fields, and schema drift in third-party responses
• Experience in regulated or compliance-adjacent domains (fintech, legaltech, healthtech) — you treat data correctness as a hard constraint, not a quality-of-life concern
• Worked on a system where workflow state durability mattered — you've thought about what happens to in-flight processes when you deploy a breaking change