Question 1

What is the difference between Array and Linked List?

Accepted Answer

Arrays store elements in contiguous memory locations with O(1) random access but O(n) insertion/deletion. Linked Lists store elements in nodes connected via pointers with O(n) access but O(1) insertion/deletion at known positions. Arrays are better for read-heavy operations and cache-friendly access. Linked Lists are better for frequent insertions/deletions and dynamic sizing. In real applications, arrays (ArrayList/Vector) are used more often due to CPU cache efficiency.

Question 2

Explain the differences between Stack and Queue.

Accepted Answer

Stack follows LIFO (Last In, First Out) — like a stack of plates. Push/pop operations happen at the top. Used in: function call management, undo operations, expression parsing, DFS. Queue follows FIFO (First In, First Out) — like a line at a counter. Enqueue at rear, dequeue from front. Used in: BFS, task scheduling, print queue, message buffers. Both have O(1) push/pop operations. Variants include Priority Queue (elements dequeued by priority) and Deque (double-ended queue).

Question 3

What is the time complexity of common sorting algorithms?

Accepted Answer

Bubble Sort: O(n²) avg/worst — educational only. Selection Sort: O(n²) — always scans full remaining array. Insertion Sort: O(n²) avg, O(n) best — great for small/nearly sorted arrays. Merge Sort: O(n log n) always, stable, uses O(n) extra space. Quick Sort: O(n log n) avg, O(n²) worst — fastest in practice, in-place. Heap Sort: O(n log n) — in-place but not stable. Tim Sort: O(n log n) — hybrid, used in Python/Java standard libraries. For interviews, focus on Merge Sort and Quick Sort implementations.

Question 4

Explain how a Hash Map works internally.

Accepted Answer

A HashMap stores key-value pairs using hashing. The key's hashCode() is computed and mapped to a bucket index (hash % array_size). Each bucket holds entries — in Java 8+, it's a linked list that converts to a red-black tree when entries exceed 8 (treeification). On collision, new entries are added to the chain/tree. Load factor (default 0.75) triggers resize/rehash when exceeded, doubling the array size. Average operations: O(1) for get/put/delete. Worst case: O(log n) with tree, O(n) with list.

Question 5

What is Dynamic Programming? Give an example.

Accepted Answer

Dynamic Programming (DP) solves complex problems by breaking them into overlapping subproblems and storing results to avoid recomputation. Two approaches: Top-down (memoization) — recursive with cache; Bottom-up (tabulation) — iterative, building from base cases. Example: Fibonacci. Naive recursion is O(2^n). With DP: dp[0]=0, dp[1]=1, dp[i]=dp[i-1]+dp[i-2] — O(n) time, O(n) space (or O(1) with two variables). Key DP problems: Knapsack, Longest Common Subsequence, Coin Change, Matrix Chain Multiplication.

Question 6

How would you design a URL shortener like bit.ly?

Accepted Answer

Requirements: Shorten URLs, redirect to original, analytics. Design: Generate unique 7-char Base62 key (62^7 = 3.5 trillion URLs). Storage: Key-value store (Redis for hot data, PostgreSQL for persistence). Key generation: Counter-based with range partition across servers, or random with collision check. Flow: POST /shorten → generate key → store mapping → return short URL. GET /:key → lookup → 301 redirect. Add: Rate limiting, analytics (click count, geo, referrer), TTL expiration, caching (90%+ cache hit rate). Scale: Read-heavy (100:1 ratio), horizontal scaling with consistent hashing.

Question 7

Design a chat application like WhatsApp.

Accepted Answer

Core features: 1-on-1 messaging, group chat, online status, read receipts. Architecture: WebSocket connections for real-time messaging (persistent connection per client). Message flow: Client → Load Balancer → WebSocket Server → Message Queue (Kafka) → Recipient's WebSocket Server → Client. Storage: Cassandra for messages (write-optimized, partitioned by chat_id), Redis for online status and last seen. Media: S3 + CDN for images/videos. Groups: Fan-out on write (small groups) vs fan-out on read (large channels). Delivery: Store-and-forward pattern, retry queue for offline users. Scale: Partition by user_id, 10M+ concurrent connections via server fleet.

Question 8

How would you design a rate limiter?

Accepted Answer

Algorithms: 1) Token Bucket — bucket fills at fixed rate, each request consumes a token; allows bursts. 2) Sliding Window Log — store timestamp of each request, count within window; precise but memory-heavy. 3) Sliding Window Counter — combine fixed window counts with weighted overlap; good balance. Implementation: Redis with INCR + EXPIRE for distributed rate limiting. Key design by user_id, IP, or API key. HTTP headers: X-RateLimit-Limit, X-RateLimit-Remaining, X-RateLimit-Reset. Return 429 Too Many Requests when exceeded. For distributed systems, use Redis Cluster with Lua scripts for atomic operations.

Question 9

Explain the four pillars of OOP with examples.

Accepted Answer

1) Encapsulation: Bundling data + methods, hiding internals. Example: BankAccount class with private balance, public deposit()/withdraw(). 2) Abstraction: Hiding complexity, showing essential features. Example: Car's accelerate() method hides engine mechanics. 3) Inheritance: Child class inherits parent's properties. Example: Dog extends Animal, gets eat() method, adds bark(). 4) Polymorphism: Same interface, different implementations. Compile-time: method overloading (multiple add() with different params). Runtime: method overriding (Shape.draw() behaves differently for Circle vs Rectangle).

Question 10

What are SOLID principles?

Accepted Answer

S — Single Responsibility: A class should have one reason to change. Don't make a User class that handles both data and email sending. O — Open/Closed: Open for extension, closed for modification. Use interfaces/abstract classes. L — Liskov Substitution: Subtypes must be substitutable for their base types without breaking behavior. I — Interface Segregation: Many specific interfaces are better than one general interface. Clients shouldn't implement methods they don't use. D — Dependency Inversion: Depend on abstractions, not concretions. Use dependency injection. These principles make code maintainable, testable, and extensible.

Question 11

What is the difference between Abstract Class and Interface?

Accepted Answer

Abstract Class: Can have both abstract and concrete methods, constructors, instance variables, and state. A class can extend only ONE abstract class (single inheritance). Use when classes share common behavior. Interface: Only abstract methods (Java 8+ allows default/static methods), no constructors, no state (only constants). A class can implement MULTIPLE interfaces. Use for defining capabilities/contracts. Example: "Animal" as abstract class (has common eat() method), "Swimmable" and "Flyable" as interfaces (capabilities). Modern practice: Prefer interfaces for loose coupling; use abstract classes when sharing code.

Question 12

What is the difference between SQL and NoSQL databases?

Accepted Answer

SQL (Relational): Structured schema, tables with rows/columns, ACID transactions, JOIN operations. Examples: PostgreSQL, MySQL, Oracle. Best for: Complex queries, relationships, data integrity (banking, e-commerce). NoSQL: Flexible schemas, various data models. Types: Document (MongoDB), Key-Value (Redis), Column-family (Cassandra), Graph (Neo4j). BASE instead of ACID. Best for: High scalability, unstructured data, rapid development. Decision factors: Data structure, query patterns, consistency vs availability (CAP theorem), scale requirements. Many modern systems use both (polyglot persistence).

Question 13

Explain database indexing and when to use it.

Accepted Answer

An index is a data structure (typically B-Tree or Hash) that speeds up data retrieval at the cost of additional storage and slower writes. Like a book's index — instead of scanning every page, you jump directly to the right section. When to use: Columns frequently used in WHERE, JOIN, ORDER BY clauses; Columns with high cardinality (many unique values). When NOT to use: Small tables, columns with low cardinality (boolean), heavily write-intensive tables, columns rarely queried. Types: Single-column, Composite (multi-column, order matters), Unique, Partial, Covering. A query using a composite index (a, b, c) can use it for (a), (a,b), or (a,b,c) — leftmost prefix rule.

Question 14

What are ACID properties in databases?

Accepted Answer

Atomicity: Transaction is all-or-nothing. If any part fails, the entire transaction is rolled back. Like a bank transfer — debit and credit must both succeed. Consistency: Transaction brings database from one valid state to another. All constraints, triggers, and rules are satisfied. Isolation: Concurrent transactions don't interfere. Isolation levels: Read Uncommitted → Read Committed → Repeatable Read → Serializable (strongest). Durability: Once committed, data persists even after system failure (write-ahead logging, checkpointing). ACID is crucial for financial systems, inventory management, and any system requiring data accuracy.

Question 15

Explain the difference between REST and GraphQL.

Accepted Answer

REST: Multiple endpoints (/users, /users/1/posts), fixed data structure per endpoint, uses HTTP verbs (GET/POST/PUT/DELETE), can over-fetch or under-fetch data, caching via HTTP (ETags, Cache-Control), simpler to implement. GraphQL: Single endpoint (/graphql), client specifies exactly what data it needs in the query, no over-fetching, strongly typed schema, introspection for documentation, real-time via subscriptions. When to use REST: Simple CRUD, caching-heavy, public APIs. When to use GraphQL: Complex relational data, mobile apps (bandwidth sensitive), rapid iteration. Many companies use both.

Question 16

What happens when you type a URL in the browser?

Accepted Answer

1) Browser checks cache (local, OS, router, ISP). 2) DNS lookup: Resolves domain to IP via recursive DNS query (root → TLD → authoritative). 3) TCP handshake: SYN → SYN-ACK → ACK (3-way). 4) TLS handshake (for HTTPS): Certificate verification, key exchange, encrypted session. 5) HTTP request: Browser sends GET request with headers. 6) Server processes: Load balancer → web server → application → database. 7) HTTP response: Server sends HTML with status code. 8) Browser rendering: Parse HTML → build DOM, parse CSS → build CSSOM, merge into Render Tree → Layout → Paint → Composite. 9) JavaScript execution: Parse, compile, execute (V8 engine). 10) Additional resources: Fetch CSS, JS, images (parallel connections, HTTP/2 multiplexing).

Question 17

What is the difference between cookies, localStorage, and sessionStorage?

Accepted Answer

Cookies: 4KB limit, sent with every HTTP request (overhead), server-accessible, can set expiry/domain/path, used for authentication (httpOnly, secure flags). localStorage: 5-10MB, persists forever until cleared, same-origin only, not sent to server, synchronous API. sessionStorage: Same as localStorage but data cleared when tab closes. Use cookies for: Auth tokens, server-side sessions, cross-sub-domain data. Use localStorage for: User preferences, cached data, non-sensitive state. Use sessionStorage for: Temporary form data, single-session state. For sensitive data, prefer httpOnly secure cookies over any client-side storage.

Software Engineering Interview Questions & Answers

Data Structures & Algorithms

System Design

Object-Oriented Programming

Database & SQL

Web Development

Key Tips for Software / IT Interviews