Rafiq. Learning with parents in control.
Rafiq creates educational games for children aged eight to ten. We built its AWS foundation and mobile apps around parent-owned profiles, dependable content delivery and a release process the team can operate as the learning library grows.
The learning experience depended on the foundation.
Rafiq needed to move from a landing page to an interactive product for children, parents and schools. Accounts, progress records, game assets and lesson videos had to work together without creating separate credentials for every child.
The platform also needed to handle after-school demand and varied connections. A small team needed a reliable way to add worlds and missions while keeping parents in control of their children's profiles.
Start with the family and build outward.
Make the parent the account owner
Children play under parent-owned profiles, with a consent flow before profile activation. Parent, teacher and school-administrator access is handled through the identity service.
Bring content closer to the player
CloudFront serves game assets and adaptive lesson video, while the apps cache assets for offline-tolerant play. A media pipeline prepares video for different connection speeds.
Separate gameplay from reporting demand
Elastic ECS Fargate services handle application traffic. Aurora supports progress records and read-heavy parent and school dashboards, with monitoring and recovery controls around the data.
Prove a complete slice with families
The proof of concept included sign-up, a child profile, one world's missions and video delivery. Pilot use and load tests informed the production build, rather than treating infrastructure sizing as a fixed guess.
A launch-ready product with a repeatable release process.
The engagement gave Rafiq a production platform and iOS and Android apps ready for public launch. Infrastructure code, deployment pipelines, runbooks and training give its engineers a way to release new content and operate the platform themselves.
The design targets 99.9% monthly availability and API latency below 300 ms at the 95th percentile under forecast peak load. These objectives describe the intended service level, not a measured public-traffic record.