ITEN

Cases

Real situations, described without names or numbers: the problem, what I did, and the method behind it.

01 A slot swap on a Function App in production

Problem: After a deployment slot swap, a connection string that was not bound to the slot moved with the swap, and production picked up the wrong configuration.
Intervention: Configuration treated as code, scripts to save and restore the App Settings, and a rollback ready before each swap.
Method: Checklist before the swap, verification after it, a written restore procedure anyone can run again.

02 APIs on a management system with several databases

Problem: The same customer had different IDs in different databases; collations and date formats did not match; some writes stopped halfway.
Intervention: An OpenAPI contract, explicit mapping between the IDs, repeatable operations and a log for every write.
Method: Real cases first, then the contract; every write can be traced and safely repeated.

03 AI in production, with human checks

Problem: AI writes code quickly, but a system in production cannot afford surprises.
Intervention: The code is written by AI too; reading the diff, the smoke test and the rollback stay human.
Method: Project rules, documentation in the repository, agents with explicit handoffs, no secrets in the agent’s context.

04 For an SME, the value is in its data history

Problem: Years of data spread across spreadsheets and old software, with nobody using it.
Intervention: Clean the data, connect the sources, then automate one process.
Method: One step at a time, measurable, before moving on to the next.