#payments

When two engineers left mid-project and the datacenter contract renewal was weeks away, the right move was a container wrapper, not a rewrite.

When the purge backlog grew faster than we could process it, the fix was not a faster purge job. It was using the database topology we already had.

When integrating multiple encryption services across acquired brands, the answer was not more code. It was an HAProxy rule.

When we discovered stored cards that could not process recurring payments after SCA went live, we had four sprints to fix it without breaking the customer experience.

Partners were spending hours preparing invoice files we already had the data for. The fix was not a better upload form.

Tiny, non-alerting spikes in TP99 graphs led to a sharding problem that only affected our most important customers.

The team wanted to batch API calls to a Korean payment processor because that is how we did it with Chase. The abstraction was wrong.