Tickr

Tickr is an offline-first Android task management app built with Kotlin and Jetpack Compose to demonstrate production-style Android architecture.

It is designed as an interview-ready project that showcases how to build a local-first app with Room as the single source of truth, reactive UI using Flow/StateFlow, and resilient background sync using WorkManager.

Tech Stack

• Kotlin
• Jetpack Compose
• MVVM
• Room
• Repository Pattern
• Flow / StateFlow
• WorkManager
• Material 3

Features

• Create, edit, complete, and delete tasks
• Filter tasks by All, Active, and Completed
• Offline-first local persistence with Room
• Reactive task streams using Flow
• Sync states for pending, syncing, synced, and failed operations
• Soft delete flow for sync-safe task deletion
• Background retryable sync using WorkManager
• Fake remote data source to simulate production-style sync behavior
• Due date picker UI for task reminders/planning

Why This Project

Tickr was built to practice and demonstrate:

• offline-first data flow
• clean Android architecture
• local database as single source of truth
• repository-driven state transitions
• background work and retry handling
• interview-friendly design decisions that are easy to explain

Architecture Overview

Tickr follows a layered architecture:

UI (Compose)
  -> ViewModel
    -> Repository
      -> Room (single source of truth)
      -> Remote sync layer
      -> WorkManager worker

Layers

• ui Compose screens, UI state models, and ViewModels
• domain Domain models and repository contracts
• data Room entities, DAO, repository implementation, mappers, and fake remote data source
• worker Background sync scheduling and worker execution

Offline-First Data Flow

Reads

1. UI collects StateFlow from the ViewModel
2. ViewModel observes tasks from the repository
3. Repository reads from Room using Flow
4. Room emits updates whenever the local database changes
5. Compose recomposes from local state

Writes

1. User creates/updates/deletes a task
2. Repository writes the change to Room immediately
3. Task is marked with a sync state such as pending create/update/delete
4. UI updates instantly from the local database
5. WorkManager later triggers background sync
6. Sync success updates the task as SYNCED or removes soft-deleted rows
7. Sync failure marks the task as failed so it can be retried

Sync Model

Tickr separates internal sync bookkeeping from UI-facing sync state.

Internal sync states

• PENDING_CREATE
• PENDING_UPDATE
• PENDING_DELETE
• SYNCING
• SYNCED
• FAILED_CREATE
• FAILED_UPDATE
• FAILED_DELETE

UI-facing sync states

• PENDING
• SYNCING
• SYNCED
• FAILED

This keeps repository logic expressive while keeping UI state simple.

Project Structure

app/src/main/java/com/rohitkhandelwal/tickr
├── core
├── data
│   ├── local
│   ├── mapper
│   ├── remote
│   └── repository
├── di
├── domain
├── ui
│   ├── navigation
│   └── screen
└── worker

Fake Remote Sync

Tickr currently uses a fake remote data source to simulate production-style sync without requiring a real backend.

You can intentionally trigger sync failures by adding:

[fail-sync]

to a task title or description. This is useful for demonstrating:

• retry behavior
• failed sync states
• local-first UX under unreliable network conditions

Running the Project

Prerequisites

• Android Studio
• JDK 11+
• Android SDK with a recent emulator or physical device

Steps

1. Clone the repository
2. Open the project in Android Studio
3. Let Gradle sync finish
4. Run the app configuration on an emulator or device

What This Project Demonstrates in Interviews

• How to design an offline-first Android app
• Why Room should be the single source of truth
• How MVVM, repository pattern, and Flow work together
• How to model sync states and retries cleanly
• When to use WorkManager for resilient background work
• How to keep architecture production-style without overengineering

Future Improvements

• Real backend integration
• Exact recurring reminders using AlarmManager
• Hilt for dependency injection
• UI tests for major flows
• Search and sorting improvements
• Multi-module modularization if the project grows

Author

Rohit Khandelwal

• GitHub: rohitk2001