Problem
SISOC covers community kitchens, field surveys, admissions, childcare centers, food-assistance and job-training programs, and a 360° view of each citizen record. It is used by staff across provinces and areas: technicians, coordinators, legal, accounting, auditors.
The public agency behind it asks for large changes, delivered fast and cheap. Cheap here means time: we are asked to shorten delivery by spending less of it studying, validating and being careful with decisions. Commitments are often made before engineering has scoped them.
And the ground keeps moving. Requirements change from one week to the next: what was agreed this week can be different the next, often after the work has already started.
Context
SISOC grew out of an existing citizen-management system. It was not built to scale: it barely did its job, and it was not stable. It had no tests, no linters, no lower environments, no pull-request culture, no controlled deploys and no documentation. Tasks lived in a WhatsApp group.
The infrastructure is limited and partly owned by another government team. SISOC also consumes external systems it does not control, like the national identity registry. And other systems consume SISOC: a ticketing system, a communications system and three field apps.
My responsibility
I coordinate priorities with the people who talk to stakeholders, and split the work across four developers and myself. I built the standards and the automations that keep them in place. I review and approve every pull request and every production deploy, and I maintain the infrastructure. I lead architecture decisions and write them down as ADRs.
The team builds features across all modules. They built the three field apps.
