Showing posts with label Android Interview. Show all posts
Showing posts with label Android Interview. Show all posts

Common Scenarios & Examples of Coroutine Memory Leaks

 What is a Memory Leak?

A memory leak occurs when memory that is no longer needed is not released because the program still holds references to it. In Android, this means the garbage collector (GC) cannot reclaim that memory, potentially leading to OutOfMemoryError and app crashes.


 How Do Memory Leaks Occur in a Coroutine?

Coroutines are lightweight threads, but if not scoped and managed correctly, they can continue running even when the component (like an Activity or ViewModel) is destroyed. This can lead to memory leaks.





 Common Scenarios & Examples of Coroutine Memory Leaks

Here are some common scenarios where memory leaks can happen with coroutines, what causes them, what happens afterward, and how to fix them:


1. Coroutine Scope Tied to a Destroyed Component

❌ Problem:

class MainActivity : AppCompatActivity() {

    override fun onCreate(savedInstanceState: Bundle?) {
        super.onCreate(savedInstanceState)
        lifecycleScope.launch {
            delay(10_000) // 10-second delay
            Log.d("Coroutine", "Finished!")
        }
    }
}

 What Happens:

Even if the user rotates the screen or navigates away, the coroutine keeps running unless cancelled. If it references this, the Activity is retained in memory.

✅ Fix:

Use a lifecycle-aware scope, like lifecycleScope, and ensure the task is cancellable:

lifecycleScope.launch {
    withTimeoutOrNull(5000) {
        delay(10_000)
    }
}

2. GlobalScope Misuse

❌ Problem:

GlobalScope.launch {
    val bitmap = loadLargeBitmap()
    imageView.setImageBitmap(bitmap)
}

 What Happens:

  • GlobalScope lives for the entire app lifecycle.

  • If imageView belongs to a destroyed Activity, it is retained in memory.

✅ Fix:

Use viewModelScope, lifecycleScope, or a custom CoroutineScope that is cancelled appropriately:

class MyViewModel : ViewModel() {
    fun loadImage() {
        viewModelScope.launch {
            val bitmap = loadLargeBitmap()
            _bitmapLiveData.value = bitmap
        }
    }
}

3. ViewModel or Activity Holds Long-lived Coroutine with UI Reference

❌ Problem:

class MyViewModel : ViewModel() {
    var activity: MainActivity? = null

    fun doWork() {
        viewModelScope.launch {
            delay(10_000)
            activity?.updateUI() // Leaks MainActivity
        }
    }
}

 What Happens:

Even after MainActivity is destroyed, the coroutine keeps a reference to it via activity.

✅ Fix:

Avoid passing UI references. Use LiveData, StateFlow, or a callback interface that is weakly referenced.


4. Job not Cancelled in onDestroy()

❌ Problem:

class MyActivity : AppCompatActivity() {
    private val job = Job()
    private val scope = CoroutineScope(Dispatchers.Main + job)

    override fun onCreate(savedInstanceState: Bundle?) {
        super.onCreate(savedInstanceState)
        scope.launch {
            fetchData()
        }
    }

    //  Missing job.cancel()
}

 What Happens:

The coroutine continues running even after the Activity is destroyed.

✅ Fix:

Cancel the job:

override fun onDestroy() {
    super.onDestroy()
    job.cancel()
}

5. Flow Collection Without Proper Scope

❌ Problem:

viewModel.dataFlow.onEach {
    textView.text = it
}.launchIn(GlobalScope) //  WRONG

 What Happens:

This will outlive the lifecycle of the view that textView belongs to.

✅ Fix:

viewModel.dataFlow
    .onEach { textView.text = it }
    .launchIn(lifecycleScope)

Or use:

lifecycleScope.launch {
    repeatOnLifecycle(Lifecycle.State.STARTED) {
        viewModel.dataFlow.collect {
            textView.text = it
        }
    }
}

 Best Practices to Avoid Coroutine Memory Leaks

 Practice  Explanation
Use viewModelScope, lifecycleScope These scopes are lifecycle-aware and cancel automatically.
Avoid GlobalScope for UI tasks GlobalScope never cancels, leading to leaks.
Always cancel custom scopes Manually call job.cancel() in onDestroy() or similar.
Avoid long-lived references to Context/UI Don’t store Activities/Views inside ViewModels or background coroutines.
Use structured concurrency Always launch coroutines inside a well-defined scope.
Make coroutines cancellable Use withTimeout, isActive, and ensure long-running tasks honor cancellation.
Use repeatOnLifecycle for Flow Ensures collection only happens during active lifecycle state.

 Use Case Example: Fetch User Profile and Update UI

❌ Incorrect (Leaky)

class ProfileActivity : AppCompatActivity() {
    override fun onCreate(savedInstanceState: Bundle?) {
        super.onCreate(savedInstanceState)
        GlobalScope.launch {
            val user = fetchUserProfile()
            runOnUiThread {
                profileTextView.text = user.name
            }
        }
    }
}

✅ Correct (Leak-safe)

class ProfileActivity : AppCompatActivity() {
    override fun onCreate(savedInstanceState: Bundle?) {
        super.onCreate(savedInstanceState)

        lifecycleScope.launch {
            val user = fetchUserProfile()
            profileTextView.text = user.name
        }
    }
}

 Summary

πŸ›‘ Don't Do This ✅ Do This Instead
Launch coroutine in GlobalScope Use viewModelScope or lifecycleScope
Hold references to Activity/View Use LiveData/StateFlow for updates
Ignore coroutine cancellation Make coroutine cancellable
Launch coroutine in onCreate w/o guard Use repeatOnLifecycle or cancel scope

πŸ“’ Feedback: Did you find this article helpful? Let me know your thoughts or suggestions for improvements! 😊 please leave a comment below. I’d love to hear from you! πŸ‘‡

Happy coding! πŸ’»✨

System Design Interviews in Android

System design interviews can be a daunting part of the interview process for Android Engineers. While the focus often leans towards architecture, performance, scalability, and reliability, many system design concepts are transferable to mobile development, especially when working with Kotlin and Jetpack Compose. In this article, we’ll explore 12 essential algorithms that play a pivotal role in system design, offering insights into how they can be used effectively in Android Kotlin Compose-based applications.


1. Bloom Filter: Reducing Costly Lookups

A simple example of a Bloom Filter to prevent unnecessary database or network lookups.

class BloomFilter(val size: Int, val hashFunctions: List<(String) -> Int>) {
    private val bitSet = BitSet(size)

    fun add(element: String) {
        hashFunctions.forEach {
            val hash = it(element) % size
            bitSet.set(hash)
        }
    }

    fun contains(element: String): Boolean {
        return hashFunctions.all {
            val hash = it(element) % size
            bitSet.get(hash)
        }
    }
}

// Usage Example
val bloomFilter = BloomFilter(
    size = 1000,
    hashFunctions = listOf(
        { input: String -> input.hashCode() },
        { input: String -> input.length.hashCode() }
    )
)

bloomFilter.add("John")
println(bloomFilter.contains("John")) // Should return true
println(bloomFilter.contains("Alice")) // Likely false

2. Geohash: Location-Based Services

Using Geohash for nearby locations.

// Example using Geohash library
import org.geohash.GeoHash
import com.google.android.gms.maps.model.LatLng

fun getNearbyGeohash(latitude: Double, longitude: Double): String {
    val geohash = GeoHash.withCharacterPrecision(latitude, longitude, 7)
    return geohash.toBase32()
}

val geohash = getNearbyGeohash(37.7749, -122.4194) // San Francisco
println("Geohash: $geohash")

3. Hyperloglog: Estimating Unique Elements

This can be implemented by tracking unique user IDs or events in a mobile app.

// Using Hyperloglog for tracking unique views
val uniqueUsers = mutableSetOf<String>()

fun addUniqueUser(userId: String) {
    uniqueUsers.add(userId)
}

fun getUniqueUserCount() = uniqueUsers.size

// Simulate adding users
addUniqueUser("user1")
addUniqueUser("user2")
addUniqueUser("user1")

println("Unique users: ${getUniqueUserCount()}")

4. Consistent Hashing: Efficient Data Distribution

A consistent hashing example to distribute tasks.

class ConsistentHashing(private val nodes: List<String>) {
    fun getNode(key: String): String {
        val hash = key.hashCode()
        val nodeIndex = Math.abs(hash % nodes.size)
        return nodes[nodeIndex]
    }
}

val nodes = listOf("Node A", "Node B", "Node C")
val consistentHash = ConsistentHashing(nodes)

println(consistentHash.getNode("user1"))  // It could print "Node B"

5. Merkle Tree: Verifying Data Integrity

Example of a Merkle Tree used for verifying data integrity.

data class MerkleNode(val hash: String, val left: MerkleNode? = null, val right: MerkleNode? = null)

fun createMerkleTree(data: List<String>): MerkleNode {
    if (data.size == 1) {
        return MerkleNode(data[0])
    }

    val mid = data.size / 2
    val left = createMerkleTree(data.subList(0, mid))
    val right = createMerkleTree(data.subList(mid, data.size))

    val combinedHash = (left.hash + right.hash).hashCode().toString()
    return MerkleNode(combinedHash, left, right)
}

val tree = createMerkleTree(listOf("A", "B", "C", "D"))
println("Root Hash: ${tree.hash}")

6. Raft Algorithm: Consensus in Distributed Databases

A simplified simulation of Raft’s consensus in Android.

// Simulate Raft leader election process
class RaftLeaderElection(val nodes: List<String>) {
    private var leader: String? = null

    fun electLeader(): String {
        leader = nodes.random()
        return leader!!
    }
}

val raft = RaftLeaderElection(listOf("Node A", "Node B", "Node C"))
println("Leader is: ${raft.electLeader()}")

7. Lossy Count: Estimating Item Frequencies

Using the Lossy Count algorithm to estimate frequencies of items.

class LossyCount(val threshold: Int) {
    private val counts = mutableMapOf<String, Int>()

    fun add(element: String) {
        counts[element] = counts.getOrDefault(element, 0) + 1
    }

    fun getFrequencies(): Map<String, Int> {
        return counts.filter { it.value >= threshold }
    }
}

val lossyCount = LossyCount(2)
lossyCount.add("Apple")
lossyCount.add("Apple")
lossyCount.add("Banana")

println(lossyCount.getFrequencies())  // Expected: {Apple=2}

8. QuadTree: Spatial Partitioning

A basic implementation of QuadTree for location-based services.

class QuadTree(val boundary: Rect, val capacity: Int) {
    private val points = mutableListOf<LatLng>()
    private var divided = false

    fun insert(point: LatLng): Boolean {
        if (!boundary.contains(point)) return false
        if (points.size < capacity) {
            points.add(point)
            return true
        }
        if (!divided) {
            subdivide()
        }
        // Insert into the appropriate quadrant
        return true
    }

    private fun subdivide() {
        divided = true
        // Divide into 4 quadrants
    }
}

data class LatLng(val latitude: Double, val longitude: Double)
data class Rect(val latMin: Double, val latMax: Double, val lonMin: Double, val lonMax: Double) {
    fun contains(point: LatLng) = point.latitude in latMin..latMax && point.longitude in lonMin..lonMax
}

val rect = Rect(37.0, 38.0, -122.5, -123.0)
val quadTree = QuadTree(rect, 2)
val point = LatLng(37.7749, -122.4194)

quadTree.insert(point)

9. Operational Transformation: Real-Time Collaboration

Basic collaboration on shared data.

// Simulate real-time text collaboration
class OperationalTransformation {
    var document = StringBuilder()

    fun applyOperation(op: String) {
        document.append(op)
    }

    fun getDocument() = document.toString()
}

val ot = OperationalTransformation()
ot.applyOperation("Hello ")
ot.applyOperation("World!")

println("Document: ${ot.getDocument()}")

10. Leaky Bucket: Rate Limiting in APIs

Simple Leaky Bucket algorithm for controlling API rate limits.

class LeakyBucket(val capacity: Int, val leakRate: Int) {
    private var waterLevel = 0

    fun addRequest() {
        if (waterLevel < capacity) {
            waterLevel++
            println("Request added. Water level: $waterLevel")
        } else {
            println("Bucket full, try again later.")
        }
    }

    fun leak() {
        if (waterLevel > 0) {
            waterLevel -= leakRate
        }
    }
}

val bucket = LeakyBucket(capacity = 5, leakRate = 1)

bucket.addRequest()  // Should succeed
bucket.addRequest()  // Should succeed
bucket.leak()  // Leaks 1 unit

11. Rsync: File Synchronization

Simplified rsync simulation for syncing files.

fun syncFiles(source: String, destination: String) {
    println("Syncing files from $source to $destination")
    // Simulate file sync
}

syncFiles("localFile", "remoteServer")

12. Ray Casting: Collision Detection

A basic example for collision detection in Android.

// Simulate ray casting for collision detection in 2D space
fun isCollision(ray: Line, objectShape: Rect): Boolean {
    return ray.intersects(objectShape)
}

data class Line(val start: Point, val end: Point) {
    fun intersects(rect: Rect): Boolean {
        // Logic to check if the line intersects the rectangle
        return true
    }
}

data class Point(val x: Int, val y: Int)
data class Rect(val x: Int, val y: Int, val width: Int, val height: Int)

val ray = Line(Point(0, 0), Point(5, 5))
val rect = Rect(2, 2, 2, 2)

println(isCollision(ray, rect))  // Will print true if there's a collision

Each of these algorithms can be adapted to Android Kotlin Compose for efficient, scalable applications, enabling you to optimize performance and user experience.

πŸ“’ Feedback: Did you find this article helpful? Let me know your thoughts or suggestions for improvements! 😊 please leave a comment below. I’d love to hear from you! πŸ‘‡

Happy coding! πŸ’»✨

Implementing an LRU Cache in Android using Kotlin and Jetpack Compose

In Android development, caching is a critical technique for storing data locally for quicker access, reducing network calls, and enhancing the user experience. One popular caching technique is the Least Recently Used (LRU) cache, which automatically evicts the least recently accessed data when the cache reaches its limit. This post explores how to implement an LRU cache in Android using Kotlin and Jetpack Compose.


What is an LRU Cache?

Least Recently Used (LRU) is a caching strategy that removes the least recently accessed items first when the cache exceeds its defined limit. It's optimal when you want to cache a limited amount of data, such as images, objects, or API responses, and ensure only frequently used items remain in memory.

Benefits:

  • Reduces memory footprint

  • Speeds up data retrieval

  • Avoids unnecessary API calls or file reads

  • Automatically manages cache eviction


Use Cases of LRU Cache in Android

Use Case Description
Image Caching Store bitmap images fetched from the network to avoid repeated downloads
Network Responses Save HTTP responses (e.g., user profile, dashboard data) to speed up load times
File or Object Caching Cache local files or data models in memory for quick reuse
Custom In-App Caching Temporary store for autocomplete suggestions, search history, etc.

 Implementing LRU Cache in Kotlin (Without External Libraries)

Step 1: Create an LRU Cache Manager

class LruCacheManager<K, V>(maxSize: Int) {
    private val cache = object : LruCache<K, V>(maxSize) {
        override fun sizeOf(key: K, value: V): Int {
            return 1 // Customize this if needed (e.g., for Bitmaps)
        }
    }

    fun put(key: K, value: V) {
        cache.put(key, value)
    }

    fun get(key: K): V? {
        return cache.get(key)
    }

    fun evictAll() {
        cache.evictAll()
    }
}

☝️ Note: This uses Android’s built-in LruCache<K, V> from android.util.LruCache.


Example: Caching User Profiles

Let’s say you’re loading user profiles from an API, and want to cache them in memory to avoid reloading them repeatedly.

Step 2: Define a simple data model

data class UserProfile(val id: Int, val name: String, val avatarUrl: String)

Step 3: Create a ViewModel using StateFlow and LruCache

class UserProfileViewModel : ViewModel() {
    private val cache = LruCacheManager<Int, UserProfile>(maxSize = 5)

    private val _userProfile = MutableStateFlow<UserProfile?>(null)
    val userProfile: StateFlow<UserProfile?> = _userProfile

    fun loadUserProfile(userId: Int) {
        cache.get(userId)?.let {
            _userProfile.value = it
            return
        }

        viewModelScope.launch {
            // Simulate network call
            delay(1000)
            val user = UserProfile(userId, "User $userId", "https://picsum.photos/200/200?random=$userId")
            cache.put(userId, user)
            _userProfile.value = user
        }
    }
}

Jetpack Compose UI

@Composable
fun UserProfileScreen(viewModel: UserProfileViewModel = viewModel()) {
    val profile by viewModel.userProfile.collectAsState()

    Column(
        modifier = Modifier.fillMaxSize().padding(16.dp),
        verticalArrangement = Arrangement.Center,
        horizontalAlignment = Alignment.CenterHorizontally
    ) {
        Button(onClick = { viewModel.loadUserProfile((1..10).random()) }) {
            Text("Load Random User")
        }

        Spacer(modifier = Modifier.height(20.dp))

        profile?.let {
            Text(text = it.name, style = MaterialTheme.typography.titleLarge)
            AsyncImage(
                model = it.avatarUrl,
                contentDescription = "Avatar",
                modifier = Modifier.size(120.dp).clip(CircleShape)
            )
        } ?: Text("No user loaded")
    }
}

Best Practices for Using LRU Cache

Tip Description
Use LruCache only for in-memory caching Do not use it for persistent or disk-based caching
Define a sensible maxSize Based on app usage pattern and available memory
For image caching, use Bitmap size as the unit Override sizeOf() method
Evict cache on memory warnings Use onTrimMemory() or onLowMemory() callbacks
Use libraries like Coil, Glide, or Picasso They offer built-in LRU support for image loading

πŸ”š Remember

The LRU caching mechanism is a simple yet powerful technique in Android development. It keeps your UI snappy and responsive by reducing network usage and data load time. While libraries like Glide manage LRU for images out of the box, creating your custom LRU cache gives you flexibility and control—especially when caching JSON responses or domain objects.

Pro Tip: Combine LruCache with StateFlow and Jetpack Compose for real-time UI updates and smoother UX.



#AndroidDevelopment #Kotlin #JetpackCompose #LRUCache #CachingInAndroid #StateFlow #PerformanceOptimization



πŸ“’ Feedback: Did you find this article helpful? Let me know your thoughts or suggestions for improvements! 😊 please leave a comment below. I’d love to hear from you! πŸ‘‡

Happy coding! πŸ’»✨



Cracking the Senior Android Engineer Interview: What to Expect from Start to Finish

Stepping into the interview process for a Senior Android Engineer role can be both exciting and challenging. Whether you’re preparing for your dream job at a big tech firm or a promising startup, understanding the structure and topics across each interview stage is crucial.

In this blog, we’ll break down the commonly asked topics from the initial recruiter round to the final bar-raiser interview, including everything from Kotlin coding to system design and behavioral assessments.


 Initial HR/Recruiter Screening

This is mostly a soft round to gauge your fit for the role and company culture.

Topics:

  • Summary of your Android development experience

  • Key projects you've worked on (apps, user base, challenges)

  • Current role, notice period, and salary expectations

  • Why you're exploring new opportunities

  • Communication skills and professional demeanor


Online Technical Coding Round

Here the focus is on core problem-solving skills using Kotlin.

Topics to Prepare:

  • Data Structures & Algorithms:

    • Arrays, LinkedLists, Trees, Graphs, HashMaps

    • Sorting, searching, recursion, backtracking

  • Kotlin-specific Concepts:

    • Coroutines (Job, SupervisorJob, Dispatchers)

    • Flow, StateFlow, and Channels

    • Lambda expressions, extension functions, and delegates

  • Concurrency & Asynchronous Programming in Android

Tools: HackerRank, Codility, or take-home assignments


System Design Interview

This round evaluates how you architect scalable, modular Android apps.

Key Focus Areas:

  • Clean Architecture (Presentation → Domain → Data layers)

  • MVVM vs MVI vs MVP — and when to choose which

  • Real-world design:

    • Offline-first apps

    • Syncing with API and caching (Room, DataStore)

    • Push notifications & background sync (WorkManager)

  • Dependency Injection (Hilt/Dagger2)

  • Multi-module project structuring

Example: “Design a banking app with authentication, balance display, and transaction history”


Android Platform & Jetpack Deep Dive

Here, expect questions on Jetpack libraries, Compose UI, and platform internals.

Topics:

  • Jetpack Compose:

    • State management

    • Recomposition and performance pitfalls

  • Lifecycle Management:

    • ViewModel, LifecycleOwner, LifecycleObserver

  • Jetpack Libraries:

    • Navigation, Room, Paging, WorkManager, DataStore

  • Security:

    • Encrypted storage, biometric authentication, Keystore

  • Accessibility:

    • Compose semantics, TalkBack support, content descriptions


Testing & Debugging

A great Senior Android Engineer writes testable and maintainable code.

What You Should Know:

  • Unit Testing with JUnit, Mockito, MockK

  • UI Testing with Espresso and Compose testing APIs

  • Integration Testing with HiltTestApplication

  • Debugging ANRs, memory leaks (LeakCanary), performance bottlenecks

  • Using tools like Crashlytics, Logcat, StrictMode


CI/CD, DevOps, and Release Management

Modern Android teams value automation and fast feedback cycles.

Topics:

  • CI/CD tools: Jenkins, GitHub Actions, Bitrise

  • Gradle optimization, build flavors, product types

  • Feature flag implementation (Gradle + Firebase Remote Config)

  • Code quality enforcement: Detekt, Lint, SonarQube

  • Secure and efficient release strategies (Play Store, Firebase App Distribution)


Behavioral & Leadership Assessment

This round checks for team collaboration, mentorship, and decision-making skills.

Example Questions:

  • Tell us about a time you led a project or mentored a junior

  • How do you resolve disagreements with product or design teams?

  • What’s your strategy for balancing tech debt vs. feature delivery?

  • How do you stay updated with evolving Android trends?

Tip: Use the STAR method (Situation, Task, Action, Result) to answer.


Bar-Raiser or VP/CTO Round

This is the make-or-break round for many companies.

Focus Areas:

  • End-to-end ownership of features and impact

  • Trade-offs made during architecture decisions

  • Innovation, optimization, or cost-saving initiatives you've led

  • Long-term vision, technical leadership, and culture fit


πŸ”š Final Thoughts

Landing a Senior Android Engineer role isn’t just about writing great Kotlin code. It’s about demonstrating architectural mastery, leadership, and a product-first mindset across every round.

Start prepping smart by:

  • Practicing DSA in Kotlin

  • Building or refactoring a multi-module Compose app

  • Designing systems (like a chat app or e-commerce app)

  • Writing testable, clean code

  • Staying up to date with Jetpack and security best practices


What Next?

Want mock interview questions, detailed Kotlin exercises, or a full Android app architecture walkthrough? Drop a comment or subscribe for more deep-dives every week.


πŸ”— Related Reads:



πŸ“’ Feedback: Did you find this article helpful? Let me know your thoughts or suggestions for improvements! 😊 please leave a comment below. I’d love to hear from you! πŸ‘‡

Happy coding! πŸ’»✨