What this mode is
Assessment
Inventory the job and write a staged plan. No migrated code in this mode.
You receive
- An inventory and a staged plan.
Limits
No implied Delivery. No universal equivalence. The public synthetic lab is not this engagement.
Francisco Mínguez · Data & AI Engineering
An example: one analytical pipeline
A recurring analytical job is hard to change. Logic sits in mixed SQL, implicit rules, or a stack the team no longer wants to operate.
A data or engineering lead who needs one job they can change without losing control of the results.
What you provide
- Access, with rights, to the code and business rules in scope.
- Authorised samples.
- A named owner.
- A target stack and tolerances written down before results.
Two modes
What this mode is
Delivery
Produce one equivalent unit. Scope is agreed on Contact, with or without a prior Assessment.
You receive
- Migrated or refactored code, equivalence tests, reconciliation, and a runbook.
Limits
No universal equivalence without an adequate oracle. No implied full-platform migration or production operation unless contracted. The public synthetic lab is not a Databricks platform and is not the employment case.
How we contract
Assessment and Delivery are separate modes. Which mode you need, and how deep the work goes, is agreed on Contact. Depth is not tied to a mode on this page.
Evidence already on the record
- Synthetic lab
Public freeze: every source row accepted or quarantined. Line totals 400.00 = 400.00. Equality is not a count of rows and is not coverage. Run 6afdf1566acc8178. Not a client result.
- Experian
This example is the shape of the Experian work: an inventory and a plan, then a reconciled delivery. The account of those processes, the four months, and the rest of the role is on the experience page. This page is not a second case.