iSmart
Industry
education · ~1,000 schools · state programme
Duration
37 wks
Team
not published
Disciplines
Consumer · Development · Infrastructure · Security · Content
Task
Four tasks in a row: stabilise the architecture, speed up releases, qualify for the state programme, hand the project over and launch two more.
Result
Commit to release in 1–6 minutes instead of hours. Selected for the state programme on the first attempt, ranked first. Two new products for the season and an in-house launch pipeline. The team works with agents on its own: most of it at step 3 of the adoption scale.
How it went
037 wks
stage 1
Architecture first aid
DevelopmentInfrastructure
Task
The project took off and the architecture started cracking: the team spent most of its time fighting fires, the rest on bugs.
Result
Architecture stabilised and adapted to the load. Several incidents closed with the team, around the clock.
stage 2 · 9 wks
Infrastructure and the release cycle
InfrastructureDevelopment
Task
Updated the Kubernetes policies, CDN and release procedures, moved the team to new development tooling. The goal: team speed and release speed.
Result
Commit to release in 1–6 minutes instead of hours. A dynamically scaling architecture with costs fitted to the school-year peaks.
stage 3 · 13 wks
Qualifying for the state programme
SecurityContentDevelopment
Task
Reinforced the team and ran a hard sprint for the application: children’s personal-data audit, the platform adapted to the programme, tools for methodologists, curriculum to state standards.
Result
Audit passed. Selected on the first attempt, ranked first.
stage 4 · 15 wks
Handover and two new launches
DevelopmentConsumer
Task
Handed the project over to the client’s product team and launched two more internal startups on a tight deadline — in time for the school season.
Result
The main project now lives with the in-house team. Two new products — in production for the season.
stage 5
What stayed after us
Development
Task
Worked inside the team, showing our tools in every meeting — no trainings.
Result
Most of the team is at step 3 of the AI-adoption scale, the rest at step 2 and moving up. An in-house pipeline for new projects is in place: not infinite, but repeatable.
A similar problem?
In 45 minutes we'll find where your bottleneck is and what we'd do about it in two weeks. Free, with no presentation about us.