The project record is the object at the centre of ReDesk. Seven products read and write it, and every guarantee we make about traceability rests on it being correct. You would own that surface — its schema, its consistency model, its public API and the events it emits.
What you would be doing
- Own the project, site and asset schemas, and the migration path when they change under live customers.
- Design the public REST endpoints and webhook contracts other teams and customers build against.
- Make concurrent writes from field, procurement and finance surfaces converge predictably, and surface conflicts instead of silently resolving them.
- Hold the line on the audit trail: every state change carries an actor, a timestamp and its evidence.
- Work directly with the field-tools and telemetry teams on what the record needs to accept next.
What we are looking for
- Five or more years building production backends where correctness mattered more than throughput.
- Deep comfort with relational data modelling, row-level security and schema evolution on live systems.
- Experience designing an API that other people had to live with for years.
- A bias for making invariants enforceable in the database rather than in application code.
- Written communication good enough that a design decision survives you being on leave.
Useful, not required
This is not a CRUD service. The record has to be defensible to a lender four years after the install, which changes almost every decision you would otherwise make quickly.
How we hire
Thirty minutes with the hiring manager. Mutual, not a screen.
Ninety minutes on a real problem from our backlog. Paid.
Two conversations with people you would work beside daily.
Within three working days of the last conversation, either way.

