Must-Know Technical Interview Questions for CSE/IT Freshers

Technical interviews for CSE and IT freshers test the same six topic areas across almost every company: data structures, algorithms, operating systems, databases, OOP, and networking.

The weighting shifts by company tier. IT services firms lean harder on aptitude and basic coding; product companies and embedded firms (Cadence, Texas Instruments, Bosch) go deeper on OS internals and system design. But the topic list itself is stable. Knowing what’s coming, and what a good answer looks like, is the prep advantage most students underuse.

What Technical Rounds Actually Test

A typical campus technical round for a Tier-2 or Tier-3 college fresher runs 45 to 90 minutes and moves through three layers:

• Conceptual questions: “What is a deadlock?” / “Explain normalization.” These test whether you understand the concept, not whether you can implement it under pressure.
• Code tracing or small implementations: “Write a binary search in any language.” “What does this snippet output?” These test translation from concept to code.
• Follow-up reasoning: “Why O(log n) and not O(n)?” / “When would you prefer a linked list over an array?” These separate candidates who memorised answers from candidates who understood them.

Most students prepare for layer one and forget layers two and three. The questions below are grouped to help with all three.

Data Structures and Algorithms

DSA is the highest-frequency topic in technical rounds. The GeeksforGeeks data structures reference covers implementation detail well; the questions here focus on concepts and reasoning.

• Q1: What is the difference between an array and a linked list?

  • Arrays store elements in contiguous memory — random access is O(1) but size is fixed at allocation.
  • Linked lists store elements in nodes connected by pointers — insertion and deletion at a known position is O(1), but random access is O(n) because you traverse from the head.
  • Common follow-up: “When would you choose one over the other?” Arrays for index-heavy operations; linked lists when insertion/deletion dominates.

• Q2: How does a hash table work?

  • A hash function maps a key to an index in an underlying array (the “bucket”).
  • Collisions — when two keys map to the same index — are resolved by chaining (linked list at each bucket) or open addressing (probe to the next empty slot).
  • Average-case lookup and insert: O(1). Worst case (all keys collide): O(n).

• Q3: What is binary search and when does it apply?

  • Binary search finds an element in a sorted array in O(log n) time by repeatedly halving the search range.
  • Prerequisite: the array must be sorted. Applying binary search to an unsorted array produces wrong answers, not just slow ones.

• Q4: What does Big-O notation measure?

  • It describes how an algorithm’s runtime or space requirement grows as input size n increases, in the worst case.
  • O(1) — constant; O(log n) — logarithmic; O(n) — linear; O(n²) — quadratic. The difference between O(n) and O(n²) is catastrophic at large n.

• Q5: What is dynamic programming? Give one example.

  • Dynamic programming solves problems by breaking them into subproblems and storing results to avoid recomputation (memoisation or tabulation).
  • Classic example: Fibonacci. Naïve recursion recomputes fib(n-2) and fib(n-1) for every call — O(2^n). DP stores computed values and reduces this to O(n).

For a set of solved DSA and aptitude problems similar to what appears in technical rounds, the Ten Technical Interview Aptitude Questions article covers tracing and reasoning problems in C.

Operating Systems and Networking

OS questions are heavier at product companies, embedded firms, and anywhere the role touches system-level code. The networking subset appears across almost all IT roles.

• Q6: What is the difference between a process and a thread?

  • A process is an independent program in execution with its own memory space.
  • A thread is a unit of execution within a process; threads in the same process share memory (heap, code, data) but have separate stacks and registers.
  • Key implication: threads are faster to create and communicate, but a crash in one thread can affect the whole process.

• Q7: What is a deadlock, and how is it prevented?

  • A deadlock occurs when two or more processes each wait for a resource held by the other — and none can proceed.
  • Four conditions must all hold (Coffman conditions): mutual exclusion, hold and wait, no preemption, circular wait.
  • Prevention breaks at least one condition — commonly by imposing a resource-ordering rule (no circular wait) or requiring processes to request all resources at once (no hold and wait).

• Q8: What is virtual memory?

  • Virtual memory lets a process address more memory than physically available by mapping virtual addresses to physical RAM and, when RAM is full, to disk (swap space).
  • The OS manages this via a page table; a page fault occurs when the accessed page is not in RAM, triggering a disk fetch.

• Q9: Describe the OSI model layers.

  • Seven layers, bottom to top: Physical, Data Link, Network, Transport, Session, Presentation, Application.
  • The Wikipedia OSI model article is the standard reference. Common interview question: “Which layer does TCP operate at?” Answer: Transport (Layer 4). “IP?” Answer: Network (Layer 3).

Databases and SQL

DB questions appear across every IT role: web development, backend services, and data engineering. The ACID-normalization-index triad is the core.

• Q10: What is normalization? Why does it matter?

  • Normalization is the process of structuring a relational database to reduce redundancy and prevent update anomalies.
  • 1NF removes repeating groups; 2NF removes partial dependencies; 3NF removes transitive dependencies. Most production databases aim for 3NF.

• Q11: What are ACID properties?

  • Per the ACID Wikipedia entry: Atomicity (transaction fully completes or fully rolls back), Consistency (data moves from one valid state to another), Isolation (concurrent transactions do not see each other’s intermediate state), Durability (committed transactions survive system failure).
  • If an interviewer asks “Why does Isolation matter?”, the answer is race conditions: two concurrent bank transfers on the same account can corrupt the balance without proper isolation.

• Q12: What is a database index, and how does it improve performance?

  • An index is a data structure (commonly a B-tree) that allows the database engine to locate rows without scanning the full table.
  • Trade-off: faster reads, slower writes (the index must be updated on every insert/update/delete). Index on frequently queried columns; avoid over-indexing write-heavy tables.

• Q13: Write an SQL query to find duplicate records in a table.

SELECT email, COUNT(*) AS occurrences
FROM users
GROUP BY email
HAVING COUNT(*) > 1;

• GROUP BY collapses rows with the same value; HAVING COUNT(*) > 1 filters to groups with more than one row. WHERE runs before aggregation; HAVING runs after.

Object-Oriented Programming

OOP questions are common at companies that use Java, Python, or C++ as their primary language. For hardware and embedded roles (such as the technical round at Cadence), expect OOP questions alongside C/C++ specifics.

• Q14: What is encapsulation?

  • Bundling data (attributes) and methods (behaviour) into a single unit (class) and restricting direct access to internal state. Achieved through access modifiers: private, protected, public.
  • Why it matters: changes to internal implementation do not break code that uses the class through its public interface.

• Q15: What is the difference between inheritance and polymorphism?

  • Inheritance: a child class acquires the properties and methods of a parent class (extends in Java, : in C++).
  • Polymorphism: a single interface has multiple implementations. A Shape reference can point to a Circle or a Rectangle and call the same area() method — each class provides its own version.

• Q16: What are the four principles of OOP?

  • Encapsulation, Abstraction, Inheritance, Polymorphism.
  • Abstraction hides implementation details behind a clean interface; Encapsulation hides internal state.

Networking Basics

• Q17: What is the difference between TCP and UDP?

  • TCP (Transmission Control Protocol) is connection-oriented: it establishes a handshake, guarantees delivery, and retransmits lost packets. Used where correctness matters: file transfers, email.
  • UDP (User Datagram Protocol) is connectionless: no handshake, no retransmission. Used where speed matters more than guaranteed delivery: video streaming, DNS, online games.

• Q18: How does HTTPS work?

  • HTTPS is HTTP over TLS (Transport Layer Security). During the TLS handshake, the server presents a certificate; the client verifies it against a trusted certificate authority; both derive a shared symmetric key for encrypting the session.
  • Symmetric encryption during data transfer; asymmetric (public-key) only during the handshake, because asymmetric is computationally expensive.

How to Prepare Across All Six Areas

A reasonable sequence for an 8-week campus placement prep cycle:

• Weeks 1–3: DSA foundations — arrays, strings, linked lists, stacks, queues, trees. Focus on time and space complexity for each.
• Weeks 4–5: OS and networking — process management, memory management, the OSI stack. Concept-first, then the “why does it matter for software” angle.
• Weeks 6–7: Databases and SQL — normalization, ACID, indexes, and 30+ SQL query exercises.
• Week 8: OOP review, mock interviews using questions like the 18 above, and company-specific preparation.

The KPMG interview process guide shows what a full technical interview cycle looks like in practice, beyond question memorisation.

Once you’ve traced through these 18 questions on paper, the next step is seeing them run in actual code. TinkerLLM lets you paste a binary search implementation and ask why the loop condition uses mid = left + (right - left) / 2 instead of (left + right) / 2. That kind of edge-case reasoning sticks far better than re-reading notes. At ₹299 on TinkerLLM, that is one reference book less.