Smart offline cache
Prioritises syllabus-aligned modules and queues uploads during connectivity windows.
Atlas Android
A Kotlin-first learning companion that works even without network.
Atlas powers content discovery, quizzes, and mentorship chat for remote learners with unreliable connectivity.
Problem · Approach
The challenge
Learners in low-connectivity areas needed frictionless access to content and study plans without depending on a stable connection.
How we solved it
Built with Jetpack Compose, Room, and WorkManager to keep content synced in the background and surface ML-driven recommendations.
Impact at a glance
Active learners
12k
Students across 3 time zones using offline-first mode weekly.
Sync success
>99%
Background sync completion thanks to WorkManager + conflict resolution.
DAU growth
62%
Increase in repeat usage within the first two release cycles.
What made the difference
ML recommendations
On-device TensorFlow Lite model suggests revision paths and mentor prompts.
Compose-first UI
Adaptive layouts with smooth transitions across phone and tablet breakpoints.
Delivery timeline
Product framing
Shadowed learners and mapped drop-off moments caused by unreliable network conditions.
Pilot launch
Released to 500 testers with feature flags controlling offline scope per cohort.
General availability
Rolled out analytics, mentor chat, and ML recommendation loops.
Results
Daily active usage increased 62% and support tickets dropped 40% after go-live thanks to predictable sync and clear progress cues.