LEEE 2026 Interview Experiences

About This Chapter

This chapter is compiled directly from the LEEE 2026 Discord server discussions (May 23-26, 2026) and personal interview notes from the batch that gave interviews on May 23-, 2026. All irrelevant banter has been filtered out; only study-relevant content is retained.

Process Overview

• Written exam qualifies you for the interview only. It is not part of the final selection score.
• Interview is the sole criterion for seat allocation.
• Results: Sent via email to all candidates - both accepted and rejected. Expected $\sim$8-10 days after interviews.
• CGPA and exam marks are used at most as a last-resort tiebreaker; they are not a primary filter.
• $\sim$60 total interviews conducted (ECD + CSD) for $\sim$30 seats (2026 batch).

Interview Format

• Multiple panels run simultaneously (Panel 3, 4, 5 were visible in 2026).
• Each panel has 2-3 professors.
• Many panels start immediately with questions - no intro is requested. Give your intro proactively if not prompted.
• Duration ranged from $\sim$4 minutes to 30+ minutes. Duration is not a reliable indicator of selection outcome.
• Panelists frequently give hints - use them, do not give up prematurely.
• Approach and reasoning matter more than the correct final answer.
• Think out loud. Narrate your reasoning process even when unsure.

Subjects Commonly Asked (2026 Batch)

Subject	Observed Frequency
Linear Algebra	Very High - asked in almost every ECE panel
DBMS	High
DSA	High (especially CSE/CSD)
Probability & Statistics	High
Operating Systems	Medium
COA	Medium
Theory of Computation	Medium
Signals & Systems	Seen in ECD panels
Digital Logic / CMOS	Seen in ECE panels
Discrete Mathematics	Seen in some panels
Machine Learning	Seen occasionally
Computer Graphics	Seen occasionally - mention only if you genuinely know it
Computer Networks	Seen occasionally
Group Theory	Some panels grilled heavily on this

Questions by Topic

Items marked were directly reported by candidates this year. Remaining items are broader prep topics for each subject area.

Linear Algebra

• Difference between scalar product and dot product (they are not the same thing).
• Eigenvectors and eigenvalues - definition, how to compute them.
• Determinant of a matrix involving eigenvectors.
• Given two vectors in different dimensions, find a third one. (Answer: not possible.)
• Topics leaned more towards matrices than vector spaces for many candidates.
• Rank of a matrix - definition and computation.
• Linear independence - definition, how to verify.
• Vector spaces and subspaces.
• Geometric interpretation of eigenvalues.
• Applications to ML: PCA (principal component analysis) and SVD (singular value decomposition).

DBMS

• DDL vs DML commands.
• What happens when you delete a record? (Primary key loss, foreign key cascades, unique data loss - e.g., if the only student in a course is deleted, that course record disappears too.)
• Normalization.
• B-Trees vs binary trees for indexing - why B-Trees? (Better time complexity; ${\log }_k$ where $k$ = branching factor, so more data fits per node reducing I/O.)
• SQL query writing.
• “What data structure gives ${\log }_3$ or ${\log }_4$ search complexity?” $\to$ B-tree.
• Normal forms (1NF, 2NF, 3NF, BCNF) - when a relation violates each.
• Clustered vs non-clustered indexes.

DSA

• Implement a stack using arrays.
• Implement a queue using arrays.
• Merge two sorted arrays.
• Implement merge sort.
• Sort all elements in a stack - no other data structure allowed. (Recursive approach: pop, sort remaining, insert in correct position.)
• Difference between a graph and a tree.
• Cycle detection in a graph - DFS or BFS? (DFS is simpler and more natural.)
• Ways to arrange/traverse a binary tree.
• Sort using queues / sort using stacks.
• Trees: BST operations, AVL rotations, heap properties and heapify.
• Linked lists: reversal, cycle detection, merge operations.
• Hashing: collision resolution strategies, load factor.
• Minimum spanning tree: Prim’s vs Kruskal’s - when to use which.
• Time complexity for standard operations on all data structures.

Algorithms

• Dijkstra’s Algorithm.
• Multiplying 2 numbers using divide and conquer.
• Quicksort - how it works, why it is called “quick”, worst-case scenario and how to avoid it.
• Prim’s and Kruskal’s - differences, proofs of correctness, when to prefer each.
• Dynamic programming: identifying subproblems, writing recurrences.
• Divide and conquer: merge sort, binary search, Master’s theorem.
• NP-completeness - what it means, common NP-complete problems.

Probability

• Monte Carlo method for $\pi$ estimation.
• 3 ants on a triangle - probability of non-collision.
• Grid probability question: 4$\times$4 grid, robot at (1,1) must reach (3,3). Moves: N=0.4, S=0.4, E=0.1, W=0.1. Can it reach the destination? If yes, minimum steps?
• Covariance and correlation.
• How to detect AI-generated student assignments using statistical methods.
• Conditional probability word problems.
• Bayes’ theorem - statement and application.
• Expected value calculations.
• Classic puzzles: Monty Hall problem, Birthday problem.
• Common distributions: Binomial, Normal (Gaussian), Poisson - properties and typical use cases.

Operating Systems

• Page faults.
• “Why do we need virtual memory if we already have RAM?”
• LRU (Least Recently Used) page replacement policy.
• Process vs thread - key differences.
• Deadlock - Coffman conditions, detection, prevention.

COA / Digital Logic

• CMOS design for NAND gate.
• Counter design using flip-flops (3-bit / 4-bit).
• Implementing logic functions using MUX.
• State machine design.
• Latch vs flip-flop - key difference.
• Boolean expression simplification using K-map.

Discrete Mathematics

• Proof of irrationality of $\sqrt{2}$ (very commonly asked).
• Graph coloring and chromatic number.
• Tree properties - number of edges, degree sum theorem.
• Induction proofs (e.g., sum of first $n$ natural numbers).
• Pigeonhole principle and its applications.
• Set theory proofs: union, intersection, complement, De Morgan’s laws.
• Propositional and predicate logic - connectives, quantifiers, truth tables.

Group Theory

• Some panels grilled heavily on group theory.
• Group axioms: closure, associativity, identity, inverse.
• Subgroups and cosets.
• Lagrange’s theorem.

Calculus / Signals

• Maxima and minima (single-variable and double-variable calculus).
• Power signals vs energy signals.
• Fourier transform of standard signals.
• Laplace transform and inverse Laplace transform.
• Convolution computation.
• Sampling theorem.

Electronic Devices (ECE Track)

• BJT/MOSFET amplifier analysis.
• Op-amp circuit analysis.

Circuit Theory (ECD Track)

• Thevenin’s and Norton’s theorems - statement and circuit reduction.
• Resonance: series and parallel LC circuits, resonant frequency.

C Programming / Systems

• Pointer arithmetic problems.
• Dynamic memory: malloc, calloc, free - when to use each.
• Recursion tracing.
• Memory layout: stack, heap, code, and data segments.

Puzzles and Lateral Thinking

• Cylindrical container full of water: reduce the volume to half without any instrument.
• How do you determine which side of a line a given point lies on?
• Given $n$ numbers, find duplicates - optimise for time and space.
• How many comparisons are needed to find a missing number in a linked list?
• How would you approach a completely new and unfamiliar problem?

Individual Interview Experiences

Candidate 1

Candidate 1's Interview

Opening: Panel asked college name, how I got in, and my cutoff. I fumbled the cutoff by 2 marks. Favourite subjects declared: DS, Probability, DBMS.

Probability:

• What probabilistic experiment is used for $\pi$ estimation? I said geometric distribution and wrongly explained it as “up to first success, helps estimate the decimals of $\pi$”; the correct answer is Monte Carlo (which was in my own paper - I did not recall it in the moment).

• AI-generated assignment detection: what statistical tool would you use? I mentioned covariance and correlation but could not elaborate confidently.

• 3 ants on a triangle - probability of non-collision. Got it eventually with some help from the panel.

DBMS:

• SQL commands: answered DDL and DML.

• What happens if you delete a record? Fumbled initially; panel hinted about primary key. Then mentioned foreign key and unique data loss.

DSA / Trees / Indexing:

• Binary tree - how many children? Answered 2.

• Binary tree log base? Answered 2.

• B-Trees or binary trees for DBMS indexing? Answered B-Trees.

• Why B-Trees? Guided towards time complexity (${\log }_2$ for binary tree, ${\log }_4$ for B-tree with 4 data per node); answered correctly, then incorrectly, then correctly again. Mentioned log laws.

• Follow-up: “What if a node holds 4 or 8 values?” - could not answer.

Ending: Panel asked “Any questions for us?” I said no because I was flustered. In hindsight, always prepare a question.

Candidate 2 - Selected

Candidate 2's Interview ( $\sim$25 min)

• Started with a probability case study combined with a networks case study - entire interview was built around this case study.
• LA section: bombed it. Could not answer some matrix-type questions they asked.
• When told to leave after LA, pushed back and said his strong subject was DBMS - panel agreed to continue. DBMS section went well.
• Panel did not ask for a self-introduction, went straight into questions. Others in the batch had briefly touched on their research/projects during their self-introduction; the end-of-interview question became the only opportunity to bring it up.
• At the end, panel asked: “Is there anything else you would like to add before we end the interview?”, used this to mention research papers (INTERSPEECH 2026 + 3 under review). Panel recognized one venue and said “we will consider this in the final selection.”
• CN was not part of the official written exam syllabus; did not think anyone in the 2026 batch was asked CN in the interview (outside of the case study context). Personally referred Forouzan for CN in college exams (questions aligned to GATE syllabus).
• Key insight: Redirecting to a strong subject when dismissed was likely a positive signal - showed initiative and self-awareness.

Candidate 3 - Selected

Candidate 3's Interview

Subjects: OS, COA, Linear Algebra.

• How to find eigenvectors and eigenvalues.
• Page fault.
• Difference between graph and tree (initially fumbled; panel gave hints and he corrected himself).
• Cycle detection in a graph: said BFS, but DFS is the simpler/standard answer.
• Could not precisely define eigenvalues; described eigenvectors as “linearly independent solutions for that eigenvalue.”
• Forgot to mention lab interests at IIITH; clarified when asked later.

Candidate 4

Candidate 4's Interview

• Subjects declared: DSA, Probability, DBMS, Digital Logic.
• Subjects actually asked: DSA and DBMS only.
• Intro mentioned hobbies and interest in CMC lab and Cybersecurity - “They grilled me hard on cybersecurity.”
• Mock interview helped significantly in staying calm.

Candidate 5

Candidate 5's Interview

• Interviewers: Zia Abbas and Aftab Hussain (recognized by voice from YouTube videos).
• Entire interview was Linear Algebra, asked by Aftab Hussain.
• Asked: difference between scalar product and dot product (they are not the same).
• Felt most answers were incorrect.
• Commented: “They called it electronics and asked only math.”

Candidate 6 - Selected

Candidate 6's Interview ( $\sim$15 min)

• Panel: 1 ma’am, 1 young North Indian professor, 1 older South Indian professor.
• Mostly asked about paper/research work.
• No Linear Algebra or Probability asked at all.
• Mentioned interest in CVIT, robotics, AI.
• Was in IROC top-25 last round.

Candidate 7

Candidate 7's Interview

• Subjects declared: TOC, DBMS, Probability.
• Actually asked: TOC (basic, easy) and Probability (harder).
• CGPA was explicitly asked.
• Panel 3.

Candidate 8

Candidate 8's Interview ( $\sim$22 min)

• Subjects: DSA, DBMS.
• No maths or CGPA asked.
• Received the 4$\times$4 grid probability question with directional move probabilities.
• Asked to implement algorithms and data structures.
• Explained thought process by writing and using hand gestures.
• Came home the night before at 3am and was sleepy throughout the interview.

Candidate 9

Candidate 9's Interview ( $\sim$30 min)

• Professor guided heavily and was supportive throughout.
• Did not answer most questions in the first attempt but reached answers with hints.
• B-tree question: “Name a data structure with ${\log }_3$ or ${\log }_4$ search complexity.” Correctly answered B-tree.
• Felt average overall but left feeling okay because of how supportive the panel was.

Candidate 10

Candidate 10's Interview ( $\sim$4.37 min)

• Duration: exactly 4 minutes 37 seconds. Felt “kicked off the Teams call.”
• Questions: intro, asked about Anurag University and its location.
• One LA question: given two vectors in different dimensions, find a third one. Correctly said: not possible.
• One Networks question: RL circuit characteristics (felt answer was close but possibly wrong).
• Forgot to mention Computer Graphics interest (had rehearsed it in mock).
• Did not explicitly mention why IIIT-H; should have done so proactively.
• Key reminder from community: duration is not a reliable indicator of selection outcome.

Candidate 11

Candidate 11's Interview

• Asked: ways to arrange/traverse a binary tree - failed.
• Asked: sort elements using queues - failed.
• Asked: sort elements using only a stack - failed to give a clean solution; gave a rough directional attempt.
• Answered heapsort correctly when asked.
• Panel was visibly interested in KV cache research mentioned by the candidate; profile was well-received overall.
• Also discussed an NFT promotion business ($\sim$$10K USD earned); told the panel earning was fun initially but studying is better long-term.
• Key quote: “They care about approach more than answer. They want you to think loud.”
• Lesson: Bombing the technical questions but having a strong research/profile angle can still leave a positive impression - keep talking through your approach even when stuck.

Candidate 12 - Selected

Candidate 12's Interview (~30 min)

• Questions from almost every topic (very broad panel).
• OS: “Why do we need virtual memory if we already have RAM?” - could not answer.
• DSA/Algorithms: Dijkstra’s Algorithm, multiply 2 numbers using divide and conquer.
• Group Theory: grilled heavily.
• LA: answered both LA questions (only 2 asked).
• OS follow-up: LRU - answered correctly.
• Mentioned Prof Kishor Kothapalli’s paper: Fast GPU algorithms for biconnected components.
• Strategy: named interest in Algorithms directly; avoided listing subjects he was not confident in.
• Mentioned wanting to join CSTAR lab.
• Froze when asked if he had questions for the panel.

Candidate 13

Candidate 13's Interview

• Digital Logic: 3 questions. Accidentally drew an AND gate instead of NAND gate; added an extra NOR at the end incorrectly.
• LA: answers unclear; panicked during follow-up.
• Panel ended with “Thanks for your time.”
• Lesson: Under pressure, double-check before drawing - NAND ≠ AND.

Candidate 14 - Selected

Candidate 14's Interview ( $\sim$20 min)

• Asked to list 1 math subject + 1 ECE subject. Was then grilled only on ECE (no math asked at all).
• One professor kept cross-questioning continuously.
• Panel ended: “We will remove you from the meeting now” + one prof jokingly said “BYE BYE.”
• Overall felt it went reasonably well.

Candidate 15 - Selected

Candidate 15's Interview ( $\sim$7 min)

• Only 2 panelists; one kept asking math questions continuously.
• First half: project work - answered well.
• Second half: DBMS + heavy Linear Algebra - weak in LA; acknowledged it.
• “Interview ended on me getting a question wrong.”
• Left with a mixed impression.

Candidate 15's Preparation Advice

• Go classes + Gate PYQs was enough for the written test, not much discrete math in the 2026 exam. Did Rosen afterwards specifically to prep for the interview.
• Started prep only 5 months before; around 3 months were serious. If you have more time, starting with Rosen from the beginning is recommended, there’s no time constraint for you.
• The 2026 exam didn’t ask much discrete math, but future years could flip the script. Be prepared for everything if you have the time.
• Only used Go classes for Discrete Math; college material was sufficient for the rest. Solved Gate PYQs throughout for all other subjects.
• If you can solve Gate PYQs, you’re on the right track. That’s really all that matters.
• Aptitude warning: Most people found aptitude very easy, but it was “almost my undoing.” If you’re not comfortable with NTSE-style aptitude, please practice, aptitude is similar to GATE aptitude style.
• The mock on the LEEE site is only to get you acquainted with the exam platform - not a difficulty indicator.
• Strategy for 2nd-year students: Prepare for LEEE alongside college - when college teaches Data Structures, study it for college and solve Gate PYQs. C programming and Engineering Math covered in 1st year for most, revise quickly and solve Gate PYQs. No need to overcomplicate it.

Candidate 16 - Selected

Candidate 16's Interview

• ECE questions: answered most.
• LA section: struggled significantly; panel provided many hints.
• Blank-stared for $\sim$2 minutes on one LA question, then eventually reached the answer.
• Immediately after answering, interviewer said: “Let’s end the interview here.”

Candidate 17 - Selected

Candidate 17's Interview (ECD/Signals)

• Calculus: maxima/minima (single-variable and double-variable).
• Signals: power signals vs energy signals.
• Got very confused and kept changing answers; felt formulas were at least correct.
• Noted: mock interview had covered more topics but actual interview drilled deep on one.

Candidate 18

Candidate 18's Interview

• Declared favourite subjects: OOPS and Discrete Mathematics.
• Actually asked: Operating Systems and Machine Learning.
• Had mentioned OS is from 4th semester and not her strongest - panel started with OS anyway.
• ML: explained only supervised/unsupervised; mentioned MATLAB vaguely.

Candidate 19

Candidate 19's Interview

• DSA: 2 basic questions (string + array). Solved both with brute force and optimal, explained time complexities.
• DBMS: 2 questions. Explained concepts well but could not write correct SQL query.
• Probability/stats: 1 basic question.
• ML: only explained supervised/unsupervised learning.
• Research interests mentioned: Robotics, Machine Learning.
• Panel seemed unsatisfied overall.

Candidate 20 - Selected

Candidate 20's Interview ( $\sim$15 min)

• Duration: $\sim$15 minutes.
• Answered all questions correctly.
• Community considered 15 minutes an average/healthy duration.

Candidate 21 - Selected

Candidate 21's Interview

• Had prepared COA, OS, and Probability thoroughly.
• Panel asked only DSA.
• Lesson: you cannot predict or control which subjects the panel will ask from.

Candidate 22 - Selected

Candidate 22's Interview

• “Interview was good” - easy questions, low cross-questioning.
• Panel did not ask “Why IIIT Hyderabad?”
• Research interests mentioned: Interpretability, Cognitive Science.
• CN was asked - 3 questions - because it was declared as a favourite subject. CN will not come up unless you specifically volunteer it as a favourite subject.
• Book recommendations for CN: Jim Kurose for CS students; William Stallings for ECE/EE students. Forouzan is more related to communication theory than computer networking.
• Advice: Mention your motivation to join IIITH and lab interests yourself, even if the panel does not ask.

Candidate 23

Candidate 23's Interview ( $\sim$30 min)

• Many questions from areas not studied.
• LA focused more on matrices than vector spaces.
• “I was butchered in the interview.”

Candidate 24 - Selected

Candidate 24's Interview

• Question asked: “How do we store lines and circles in computer graphics?”
• Context: end-semester exams were running simultaneously.

Candidate 25

Candidate 26's Interview ( $\sim$17 min)

• 6-7 questions total covering DBMS, Data Structures, and Probability Theory.
• Panel jumped directly into technical questions - no introduction, college background, or research interest was asked.
• Candidate answered all questions with reasonable confidence.

Candidate 26

Candidate 26's Interview

• Panel asked about JEE rank.
• Overall impression: okayish.

Candidate 27 - Selected

Candidate 27's Interview

• Semi-fumbled one VLSI question and one Linear Algebra question.
• Recovered well and answered a solid number of questions after the fumbles.
• Uncertain about outcome.

Candidate 28

Candidate 28's Interview

• Self-assessment: somewhere between 50-65%; overall felt weird.
• First question answered almost correctly.
• Talked about research topic - panel said they were unfamiliar with it.
• Got asked a coding question on Dynamic Programming.
• A few more questions followed; went okayish overall.

Candidate 29

Candidate 29's Interview

Questions asked:

1. What is quantum computing?
2. What are unitary matrices? What is the significance of them?
3. What is the rank of a matrix?
4. Can a matrix have a different number of independent columns and rows?
5. How do you store a graph?
6. Using a matrix, how do you find a path of length 3?
7. How do you find that a graph has more than one component using only matrices, without DFS and BFS?
8. What is the worst number of steps required to do the above?

Candidate 30 - Selected (ECD)

Candidate 30's Interview - ECD 2026, Selected, NIT ECE, CGPA 7.8

• Format: Online (Microsoft Teams).
• Asked to choose two subjects: one from Electronics, one from Mathematics.
• Chose: Signals & Systems (Electronics) and Calculus (Mathematics).
• No projects or research papers at the time of interview.

Signals & Systems questions (Electronics Professor):

• Fourier Transform of a sine wave.
• “Now what if x(t) = 0 for t < 0?” (causal sine wave) - answered incorrectly at first, self-corrected after pausing. Professor did not penalise the correction.
• Relation between convolution in the frequency domain → Modulation Theorem. Wrote it out on paper and demonstrated.
• FT of a sine wave from t = 0 to t = T. (Clicked immediately after the modulation theorem discussion.)
• Conceptual/analytical question: ant walking on a circle.

Calculus questions (a separate Math Professor joined):

• Gradient and directional derivative.
• How to find maxima/minima of a multivariable function. Initially said Jacobian - self-corrected to Hessian matrix (rt − s² criterion).

Other elements:

• Asked about research interests: mentioned superconducting quantum qubits at CVEST / CQST labs.
• Asked about CGPA. Did not ask about projects or research papers.

Candidate’s note: “They just want to check how you work under pressure and your thought process.” Selected; chose not to join IIITH.

Interview Question Bank - 2026

Note: The following questions are drawn from mock interviews, previous candidates, and preparation materials - they are not a prediction of what you will be asked, nor a reflection of actual interview difficulty, but a tool to help you gauge how firmly you stand on the fundamentals.

Introduction, Motivation & Research

Personal & Academic Motivation

1. Tell us about yourself.
2. Why do you want to join IIIT Hyderabad?
3. What are your research interests?
4. Which professor’s work did you explore?
5. Which papers did you read?
6. What subjects interest you the most?
7. Why are you interested in AI/ML?
8. Why do you want research-oriented education?
9. Why should we select you?
10. What projects are you most proud of?

Research Grilling Questions

1. Why not placements directly?
2. Why IIITH specifically?
3. Why research and not development?
4. What problem would you solve if given unlimited funding?
5. Which professor’s work interests you?
6. What area do you want to specialize in?

Stress-Test Questions

1. What if your research fails for 2 years?
2. Tell us something you learned completely on your own.
3. Explain a technical concept to a school student.
4. What is the biggest limitation of current AI?

C Programming

Basics of C

1. Why is C called a middle-level language?
2. Difference between compiled and interpreted languages.
3. What happens during compilation?
4. What is preprocessing?
5. Difference between declaration and definition.
6. What is a header file?
7. Why are header guards needed?
8. What is a linker?
9. What is an object file?
10. Why is C fast?

Variables & Memory

1. Difference between stack and heap memory.
2. What is the memory layout of a C program?
3. Global vs local variables.
4. Static vs automatic storage.
5. What is scope?
6. What is the lifetime of a variable?
7. Why are local variables stored on the stack?
8. Why is heap allocation slower?
9. What is a dangling pointer?
10. What is a memory leak?

Pointers

1. What is a pointer?
2. Why are pointers powerful?
3. Difference between pointer and variable.
4. Pointer arithmetic intuition.
5. Why does the array name behave like a pointer?
6. Difference between arr and &arr.
7. What is a double pointer?
8. Pointer-to-pointer use case.
9. Void pointer meaning.
10. Why does pointer type matter?

Tricky Pointer Questions

1. Difference between int *p; and int (*p)[10];.
2. Why does pointer arithmetic depend on datatype?
3. What happens when dereferencing NULL?
4. Difference between char *s = "hello"; and char s[] = "hello";.
5. Why cannot arrays be assigned?
6. Can a pointer point to freed memory?
7. What is a wild pointer?
8. Why are function pointers useful?
9. What is pointer aliasing?
10. Why are pointers dangerous?

Arrays & Strings

1. Difference between array and pointer.
2. Why is array size known at compile time traditionally?
3. Why does array indexing start from 0?
4. Multi-dimensional arrays in memory.
5. Why do strings end with ”0’`.
6. What happens if null terminator is missing?
7. Difference between sizeof(arr) and sizeof(ptr).
8. Why do character arrays differ from string literals?
9. Why is buffer overflow dangerous?
10. Why do arrays decay into pointers?

Functions

1. Call by value vs call by reference.
2. Why does C technically support only call by value?
3. What does recursion internally use?
4. Stack frame meaning.
5. What happens during a function call?
6. What is an inline function?
7. Why can recursion cause stack overflow?
8. What is tail recursion?
9. Function pointer usage.
10. Why do callbacks use function pointers?

Dynamic Memory Allocation

1. Difference between malloc and calloc.
2. Why does calloc initialize memory?
3. realloc intuition.
4. What happens if malloc fails?
5. Why is free important?
6. Double free problem.
7. Use-after-free meaning.
8. Why does fragmentation occur?
9. Why does heap memory persist after function returns?
10. Why is malloc slower than stack allocation?

Structures & Unions

1. Difference between structure and union.
2. Why does a union share memory?
3. Structure padding meaning.
4. Why is padding added?
5. Difference between . and ->.
6. Nested structures.
7. Self-referential structures.
8. Why do linked lists use structures?
9. typedef purpose.
10. Memory alignment intuition.

File Handling

1. What is a FILE pointer?
2. fopen modes.
3. Difference between text and binary files.
4. EOF meaning.
5. Why does buffering improve performance?
6. What happens internally during file write?
7. Why is flushing important?
8. Difference between fprintf and fwrite.
9. Why is random file access useful?
10. What is a seek operation?

Preprocessor & Macros

1. What is a macro?
2. Difference between macro and function.
3. Why are macros dangerous?
4. What is conditional compilation?
5. Why are include guards needed?
6. What does #define actually do?
7. Macro expansion pitfalls.
8. Why is inline safer sometimes?
9. What is token pasting?
10. Why do preprocessors exist?

Undefined Behavior & Traps

1. What is undefined behavior?
2. Why is i = i++; problematic?
3. Difference between ++i and i++.
4. Why is accessing freed memory dangerous?
5. Why is buffer overflow dangerous?
6. What happens if an array is accessed out-of-bounds?
7. Why is modifying a string literal dangerous?
8. Why is integer overflow tricky?
9. Sequence point meaning.
10. Why does UB allow compiler optimizations?

OS / System-Level C Questions

1. Why is C used for OS development?
2. Why are pointers important in OS?
3. Difference between user space and kernel space.
4. What is a system call?
5. Why is direct memory access dangerous?
6. Why do segmentation faults occur?
7. Stack overflow vs heap overflow.
8. Why is context switch expensive?
9. Why does C give manual memory control?
10. Why are low-level languages still important?

Very Tricky Oral Questions

1. Why are arrays and pointers not exactly the same?
2. Can a function return a pointer to a local variable?
3. Why are string literals often stored in read-only memory?
4. Why does sizeof(arr) differ inside a function?
5. Why does recursion internally need a stack?
6. Why does *(p + i) work?
7. Why does pointer arithmetic scale automatically?
8. Can NULL ever be dereferenced safely?
9. Why are dangling pointers dangerous?
10. Why does C trust the programmer so much?

Practical Pointer Tracing

int a[5] = 1, 2, 3, 4, 5;
printf("

1. What is the output?
2. Difference between int *p[10] and int (*p)[10].

Recursion Tracing

1. Trace recursive factorial for $n=4$.
2. Fibonacci recursion time complexity and why.
3. What is tail recursion and why does it matter?

Static Variable Tracing

static int x = 0;
void func() 
 x++;
 printf("

1. Predict output for 3 calls to func.
2. Explain static variable lifetime.

Probability & Statistics

Basic Probability

1. What is probability?
2. Why must probability lie between 0 and 1?
3. Why does total probability equal 1?
4. What is a sample space?
5. What is an event?
6. Mutually exclusive vs independent events.
7. Can mutually exclusive events be independent?
8. Conditional probability intuition.
9. Law of total probability.
10. What does “random variable” mean?

PMF / PDF / CDF

1. Difference between PMF, PDF, and CDF.
2. Can PMF be greater than 1? (No - PMF gives probabilities; each value $\le 1$ and sum $=1$.)
3. Can PDF be greater than 1? (Yes - PDF is a density, not a probability. The area under PDF $=1$, but the height can exceed 1.)
4. Why does the CDF never decrease?
5. Why does the CDF range from 0 to 1?
6. Difference between discrete and continuous distributions.
7. Why is probability at a single point zero in continuous distributions?
8. What does density mean in PDF?
9. Why does the integral of PDF equal 1?
10. Why does the derivative of CDF give PDF?
11. Can CDF have jumps?

Bayes Theorem

1. State Bayes’ theorem.
2. Prior vs posterior probability.
3. What is likelihood?
4. Why is Bayes’ theorem useful?
5. Medical testing false-positive intuition.
6. Why does the base-rate fallacy happen?
7. Why does Naive Bayes work surprisingly well?
8. What assumption does Naive Bayes make?
9. Why is the independence assumption unrealistic?
10. Why does Naive Bayes still perform well?

Expectation & Variance

1. What is expectation intuitively?
2. Why is expectation a “weighted average”?
3. Linearity of expectation.
4. Why does linearity hold even without independence?
5. What is variance intuitively?
6. Can variance be negative? (No. $\text{Var}(X)=E[(X-E[X]{)}^2]$ is a sum of squares $\ge 0$. It equals zero only when $X$ is a constant.)
7. Standard deviation vs variance.
8. Covariance intuition.
9. Correlation vs covariance.
10. Why does variance measure spread?

Common Distributions

1. Bernoulli distribution intuition.
2. Binomial distribution meaning.
3. Poisson distribution use cases.
4. Why does Poisson model rare events?
5. Gaussian distribution intuition.
6. Why does the normal distribution appear everywhere?
7. Central Limit Theorem intuition.
8. Uniform distribution meaning.
9. Exponential distribution memoryless property.
10. Geometric distribution intuition.
11. If you toss a coin 50 times and get heads each time, what is the probability of heads on the 51st toss? ($\frac{1}{2}$ - assuming a fair coin. Each toss is independent; past outcomes do not affect future ones (memoryless / i.i.d. assumption). This is a classic base-rate fallacy trap.)

Tricky Oral Questions

1. Can PDF be greater than 1?
2. Can variance be negative?
3. Can covariance be zero but dependence exist?
4. Does zero correlation imply independence?
5. Can two events be mutually exclusive and independent?
6. Can expectation lie outside possible values?
7. Why is probability at an exact point zero in continuous variables?
8. Can CDF decrease?
9. Why is conditional probability asymmetric?
10. Can an impossible event become possible after conditioning?

”Why?” Questions Professors Love

1. Why do probabilities sum to 1?
2. Why can variance not be negative?
3. Why must CDF be monotonic?
4. Why does the Gaussian appear naturally?
5. Why does conditioning change dependencies?
6. Why is Bayes’ theorem powerful?
7. Why is expectation linear?
8. Why does correlation differ from causation?
9. Why does Poisson model rare events?
10. Why is CLT important?

Probability Practical Questions

1. Monty Hall problem - explain intuitively.
2. Expected value of a fair die throw.
3. Given $P(A)=0.6$, $P(B)=0.4$, $P(A\cap B)=0.2$ - are $A$ and $B$ independent?
4. What is the probability of exactly 2 heads in 3 fair coin tosses?
5. Explain Bayes’ theorem: $P(A\mid B)=\frac{P(B\mid A)P(A)}{P(B)}$.
6. Difference between PDF and PMF.
7. Explain Gaussian distribution intuitively.

OOP

OOP Fundamentals

1. What is Object-Oriented Programming?
2. Difference between class and object.
3. What is encapsulation?
4. What is abstraction?
5. Difference between abstraction and encapsulation.
6. What is inheritance?
7. What is polymorphism?
8. Compile-time vs runtime polymorphism.
9. What is dynamic dispatch?
10. Why did OOP become popular?

Classes & Objects

1. What is a constructor?
2. Constructor vs destructor.
3. Why can a constructor not have a return type?
4. What is a copy constructor?
5. Difference between shallow copy and deep copy.
6. What is object lifecycle?
7. Why is the this pointer/reference needed?
8. Static member vs instance member.
9. Can an object exist without a class?
10. Why do classes improve modularity?

Inheritance

1. Why is inheritance useful?
2. IS-A vs HAS-A relationship.
3. Composition vs inheritance.
4. Why is composition often preferred?
5. Single vs multiple inheritance.
6. Diamond problem.
7. Why is multiple inheritance dangerous?
8. Multilevel, hierarchical, and hybrid inheritance.

Polymorphism

1. What is polymorphism?
2. Function overloading vs overriding.
3. Compile-time vs runtime polymorphism.
4. Why do virtual functions need a vtable?
5. Can static methods be overridden?
6. Why is operator overloading useful?

Advanced OOP Concepts

1. What is dependency injection?
2. What are design patterns?
3. Singleton pattern pros/cons.
4. SOLID principles.
5. Why is tight coupling bad?
6. Why are friend functions considered encapsulation-breaking?

Linear Algebra

Basis, Span & Linear Independence

1. What is span?
2. What is a basis?
3. Can a spanning set be linearly dependent?
4. Why must basis vectors be independent?
5. Why does a basis give a unique representation?
6. What is the dimension of a vector space?
7. What is linear independence / dependence?
8. Geometric interpretation of dependence.
9. How to test independence?
10. Why does a zero determinant imply dependence?

Matrices

1. What does a matrix represent geometrically?
2. What is the rank of a matrix?
3. Difference between row space and column space.
4. What is the null space?
5. Rank-nullity theorem intuition.
6. What makes a matrix invertible?
7. Why does a zero determinant mean singular matrix?

Eigenvalues & Eigenvectors

1. What is an eigenvector? What is an eigenvalue?
2. Geometric interpretation of eigenvectors.
3. How do you find eigenvectors of a matrix? Justify each step.

• Solve $\det (A-\lambda I)=0$ for eigenvalues $\lambda$. (Rationale: $Av=\lambda v\Rightarrow (A-\lambda I)v=0$ has non-trivial solutions iff the matrix is singular.)
• For each $\lambda$, solve $(A-\lambda I)v=0$ for the null space. (Rationale: those null vectors are exactly the eigenvectors.)

4. Can eigenvectors of a matrix be rotated as a result of multiplication? (No - by definition, $Av=\lambda v$ means $A$ only scales $v$; it does not rotate it. Non-eigenvectors are generally both scaled and rotated.)
5. Can an eigenvalue be zero? What does it imply?
6. Can eigenvalues be negative? Can a matrix have no real eigenvalues?
7. Why are eigenvectors important in ML (PCA, covariance matrices)?

Orthogonality & Determinants

1. What is orthogonality?
2. What is an orthonormal basis?
3. Dot product geometric meaning.
4. Gram-Schmidt process intuition.
5. What does the determinant represent geometrically?
6. Why does a zero determinant imply singularity?
7. Why does the determinant equal the product of eigenvalues?

PCA / ML-Oriented Questions

1. Why does PCA use eigenvectors?
2. Why is the covariance matrix symmetric?
3. Why do eigenvectors capture variance directions?
4. Difference between PCA and SVD.
5. Curse of dimensionality intuition.
6. Why does cosine similarity use the dot product?

Practical Eigenvalue Question

Find eigenvalues of $A=\left(\begin{array}{cc}2& 1\\ 1& 2\end{array}\right)$.

1. What do the eigenvalues represent geometrically for this matrix?
2. Why are eigenvectors important in ML?
3. Difference between rank and determinant.

Calculus

Limits & Continuity

1. What is a limit intuitively?
2. When does a limit fail to exist?
3. What is a continuous function? Mathematically define it. ${\lim }_{x\to c}f(x)=f(c)$ (Three conditions: (i) $f(c)$ is defined, (ii) the limit exists, (iii) they are equal.)
4. Difference between continuity and differentiability.
5. Why does differentiability imply continuity?
6. Can continuity exist without differentiability? Give an example.
7. What is a removable discontinuity?

Differentiation

1. What is a differentiable function? Mathematically define it. $f^{\prime }(x)={\lim }_{h\to 0}\frac{f(x+h)-f(x)}{h}\text{ exists (and is finite.)}$
2. Geometric meaning of derivative.
3. Why do corner points break differentiability?
4. an example of a continuous but not differentiable function other than $|x|$.

• $f(x)=x^{2/3}$ at $x=0$ (cusp; derivative blows up).
• $f(x)=x\sin (1/x)$ for $x\ne 0$, $f(0)=0$ (oscillates too wildly near 0).
• Weierstrass function (continuous everywhere, differentiable nowhere).

5. Difference between partial and ordinary derivative.
6. Chain rule intuition.

Applications of Derivatives

1. Why is a derivative useful in optimization?
2. Why is the derivative zero at extrema?
3. Second derivative interpretation (concavity).
4. Why does the gradient point in the direction of steepest ascent?
5. Gradient descent intuition.

Integration & Multivariable Calculus

1. What is integration intuitively?
2. Fundamental theorem of calculus.
3. What is a partial derivative?
4. Gradient interpretation.
5. Hessian matrix meaning and why it helps optimization.

Series & Taylor Expansion

1. What is a Taylor series?
2. Why are Taylor expansions useful?
3. Why does the exponential function appear naturally?

Calculus Practical Questions

1. Differentiate $y=x^x$.
2. Find maxima/minima of $f(x)=x^3-6x^2+9x+1$.
3. Explain gradient descent mathematically.
4. Why does the derivative represent slope?

Data Structures (DSA)

Arrays, Linked Lists, Stack, Queue

1. Why does an array support $O(1)$ random access?
2. Why is insertion in the middle expensive?
3. Why are arrays cache-friendly?
4. Singly vs doubly linked list - why is searching slow?
5. Why are linked lists cache-unfriendly?
6. Detect cycle in linked list (Floyd’s algorithm) - why does it work?
7. Why is stack LIFO? Real-world uses.
8. Implement queue using two stacks.
9. Why does BFS naturally use a queue?

Hash Tables

1. Why is hashmap average complexity $O(1)$?
2. Why does worst case become $O(n)$?
3. Chaining vs open addressing.
4. What is load factor and why does resizing help?

Trees & Heaps

1. Binary tree vs BST.
2. Worst case of BST and how AVL/Red-Black trees fix it.
3. Why is heap insertion $O(\log n)$?
4. Why is heapify $O(n)$ and not $O(n\log n)$?
5. Heap vs sorted array for priority queue.
6. Construct BST from: 10, 5, 1, 7, 40, 50. Give traversals.

Graphs

1. Difference between DFS and BFS.
2. Why does BFS guarantee shortest path in unweighted graph?
3. Adjacency matrix vs adjacency list.
4. Design LRU cache - what data structures and why?
5. What data structure for autocomplete?

Algorithms (DAA)

Sorting

1. Why is merge sort stable?
2. Worst case of quicksort and how to avoid it.
3. Why is heap sort not stable?
4. Which sorting algorithm for: nearly sorted array / huge external data / integers in small range?

Graph Algorithms

1. Why does Dijkstra fail for negative weights? Give a counterexample.
2. Bellman-Ford vs Dijkstra.
3. Prim vs Kruskal for MST.
4. Why does union-find help in Kruskal’s algorithm?

Dynamic Programming & Greedy

1. When should DP be used? Overlapping subproblems + optimal substructure.
2. Memoization vs tabulation.
3. Coin change: when does greedy fail?
4. Fractional knapsack vs 0/1 knapsack.

Complexity Analysis

1. Big-O vs Theta vs Omega.
2. Why ignore constants in asymptotic analysis?
3. Analyze: for(i=1; i<n; i*=2).
4. Solve recurrence $T(n)=2T(n/2)+n$.
5. What is P vs NP? Why does it matter?

Discrete Mathematics

Functions & Relations

1. Difference between relation and function.
2. What makes a function injective / surjective / bijective?
3. Why must an inverse require bijection?
4. Can order of quantifiers matter? Give example.
5. Pigeonhole principle - real-life example.
6. Inclusion-exclusion principle.

Set Theory & Logic

1. De Morgan’s laws. $\overline{A\cup B}=\bar{A}\cap \bar{B}$.
2. Why are real numbers uncountable?
3. Contrapositive vs converse - which is equivalent to the original?
4. Prove $\sqrt{2}$ is irrational.
5. Prove by induction: $1+2+\cdots +n=\frac{n(n+1)}{2}$.

Graph Theory

1. Degree sum formula: $\sum \mathrm{deg}(v)=2|E|$.
2. Why do trees have $n-1$ edges?
3. Euler path conditions.
4. Bipartite graph - why does an odd cycle break bipartiteness?

Digital Logic

Boolean Algebra & Gates

1. Why are NAND and NOR universal gates?
2. XOR relation: $A\oplus B=A^{\prime }B+A{B}^{\prime }$.
3. Can XOR be used to make NAND? ( directly with XOR alone - XOR is not functionally complete. But XOR with a constant can emulate NOT: $A\oplus 1=\bar{A}$. Combined with AND (itself not buildable from XOR alone), NAND could be formed. In practice, NAND cannot be built using only XOR gates.)
4. Why is XOR useful for parity checking?
5. De Morgan’s law: $(AB{)}^{\prime }=A^{\prime }+B^{\prime }$.

Signed Numbers & Overflow

1. What are signed and unsigned numbers? (Unsigned: all bits represent magnitude (0 to $2^n-1$). Signed: MSB is sign bit; two’s complement used in practice.)
2. Two’s complement: $-x=\bar{x}+1$.
3. What is overflow? When does it occur? How to detect/correct it? (Overflow occurs when the result of an arithmetic operation exceeds the representable range. For signed addition, detect when two numbers of the same sign produce a result of the opposite sign (or equivalently, when carry into the MSB $\ne$ carry out of the MSB). Correct by widening bit-width or using saturation arithmetic.)
4. Range of $n$-bit signed integer: $-2^{n-1}$ to $2^{n-1}-1$.
5. Why is subtraction implemented using addition (via two’s complement)?

Combinational Circuits

1. Half adder vs full adder. Full adder: $S=A\oplus B\oplus C_{in}$.
2. Ripple carry vs carry-lookahead adder - why does carry propagation cause delay?
3. Multiplexer, demultiplexer, encoder, decoder intuitions.
4. K-map and Boolean minimization - why are fewer gates desirable?

Sequential Circuits & Flip-Flops

1. Latch vs Flip-Flop. (A latch is level-triggered (output changes while enable is active). A flip-flop is edge-triggered (output changes only on a clock edge). Flip-flops are more predictable in synchronous design.)
2. SR latch vs D flip-flop.
3. Draw a JK flip-flop using a D flip-flop. ($D=J{Q}^{\prime }+K^{\prime }Q$. Connect $J$, $K$, and current output $Q$ through this combinational logic to the $D$ input of a D flip-flop.)
4. What is a race condition? Setup and hold time.
5. Why does metastability occur?

Counters & Registers

1. Difference between ring counter and Johnson counter. Draw if possible. Which has more cardinality (distinct states)?

• Ring counter: A single 1 (or 0) circulates through $n$ flip-flops. Distinct states $=n$.
• Johnson counter (twisted ring): The complemented output of the last FF feeds back to the first. Distinct states $=2n$.
• Johnson counter has more distinct states for the same number of flip-flops.

2. Synchronous vs asynchronous counters - why are asynchronous counters slower?
3. Finite state machine intuition.

Memory

1. SRAM vs DRAM - why is SRAM faster?
2. Why does DRAM need refresh?
3. Cache memory intuition and why it uses SRAM.

Operating Systems & Computer Architecture

Processes & Threads

1. Difference between process and thread.
2. Why are threads lighter than processes?
3. Why is context switching expensive?
4. What is thread synchronization?
5. Can threads run truly in parallel always?

CPU Scheduling

1. FCFS vs SJF - why can SJF starve processes?
2. Round Robin and why quantum size matters.
3. Preemptive vs non-preemptive scheduling.
4. What is starvation vs deadlock vs livelock?

Synchronization & Deadlock

1. What is a race condition? What is a critical section?
2. Semaphore vs mutex. Binary vs counting semaphore.
3. Coffman’s four conditions for deadlock.
4. Banker’s algorithm - safe vs unsafe state.
5. Deadlock prevention vs avoidance vs detection.

Paging, Virtual Memory & Cache

1. What is virtual memory? What is paging?
2. What is a page fault? What is thrashing?
3. FIFO vs LRU page replacement.
4. Temporal vs spatial locality.
5. Why do multiple cache levels exist?

Theory of Computation (ToC)

Finite Automata (FA)

1. What is automaton?
2. DFA vs NFA
3. Why are DFA and NFA equivalent?
4. Can NFA be more powerful than DFA?
5. Why DFA has exactly one transition?
6. What is epsilon transition?
7. Why epsilon transitions useful?
8. How convert NFA to DFA?
9. What is state minimization?
10. Why minimal DFA unique?

DFA transition:

(q,a)=q’

Regular Languages

1. What is regular language?
2. Closure properties of regular languages
3. Why regular languages closed under union?
4. Why regular languages closed under complement?
5. Why some languages not regular?
6. Pumping lemma intuition
7. Why pumping lemma used for proving non-regularity?
8. Difference between regex and DFA
9. Why finite automata have limited memory?
10. Can DFA recognize balanced parentheses?

Pumping Lemma

1. State pumping lemma
2. Why pumping lemma works?
3. How prove language not regular?
4. Why choice of string important?
5. What is pumping length?
6. Why pumping lemma only necessary condition?
7. Can pumping lemma prove language regular?
8. Common mistakes in pumping lemma proofs
9. Why infinite regular languages pump?
10. Why finite automata force repetition?

Pumping lemma structure:

s=xyz

Context-Free Grammars (CFG)

1. What is CFG?
2. Terminal vs non-terminal
3. What is derivation?
4. Leftmost vs rightmost derivation
5. Parse tree intuition
6. Ambiguous grammar meaning
7. Why ambiguity problematic?
8. Can every CFG be converted to CNF?
9. What is Chomsky Normal Form?
10. Why CFGs more powerful than regular languages?

Pushdown Automata (PDA)

1. Why PDA stronger than DFA?
2. Why stack increases power?
3. What languages PDA recognizes?
4. Difference between PDA and CFG
5. Why balanced parentheses need stack?
6. Deterministic vs nondeterministic PDA
7. Why PDA useful in parsing?
8. Can PDA recognize all CFLs deterministically?
9. Stack operation intuition
10. Why finite automata fail for nested structures?

Turing Machines

1. What is Turing machine?
2. Why Turing machine powerful?
3. Difference between DFA and TM
4. What is tape in TM?
5. Why infinite tape important?
6. Deterministic vs nondeterministic TM
7. Universal Turing machine intuition
8. What is Church-Turing thesis?
9. Why TM models computation?
10. Can TM solve every problem?

TM transition:

(q,a)=(q’,b,L/R)

Decidability & Halting Problem

1. What is decidable language?
2. What is undecidable problem?
3. What is recognizer?
4. Difference between decidable and recognizable
5. What is halting problem?
6. Why halting problem undecidable?
7. What is reduction?
8. Why reductions useful?
9. Rice’s theorem intuition
10. Can undecidable problems be approximated?

Chomsky Hierarchy

1. What are four levels of Chomsky hierarchy?
2. Difference between regular and context-free languages
3. Context-sensitive vs unrestricted grammars
4. Why hierarchy important?
5. Why more computational power means more expressive languages?
6. Type-0 grammar meaning
7. Which automata correspond to each hierarchy level?
8. Why regular languages weakest?
9. Why unrestricted grammars most powerful?
10. Why PDA sits between DFA and TM?

Computability & Complexity

1. Difference between computability and complexity
2. P vs NP
3. NP-hard vs NP-complete
4. What makes verification easier?
5. Polynomial-time reduction
6. Why SAT important?
7. Why NP-completeness matters?
8. Can NP-hard problems have polynomial solutions?
9. Why brute force often exponential?
10. Difference between solvable and efficiently solvable

Parsing & Compilers

1. Why parsing needed?
2. Lexical analysis vs syntax analysis
3. Why regex used in lexers?
4. Why CFG used in parsers?
5. LL vs LR parsing
6. Shift-reduce parsing intuition
7. Why ambiguity dangerous in programming languages?
8. Operator precedence parsing
9. Why compilers need automata theory?
10. Why recursive descent parsing works?

Very Tricky Oral Questions

1. Can DFA recognize palindrome?
2. Why balanced parentheses not regular?
3. Why stack changes computational power?
4. Can every CFL be regular?
5. Can every regular language be context-free?
6. Why pumping lemma cannot prove regularity?
7. Can TM simulate PDA?
8. Why finite automata cannot count arbitrarily?
9. Can undecidable problems exist in real life?
10. Why halting problem fundamental?

Counterexample-Based Questions

1. Give non-regular language
2. Give CFL that is not regular
3. Give ambiguous grammar
4. Give undecidable problem
5. Give deterministic CFL
6. Give non-deterministic CFL
7. Give regular language
8. Give language needing stack memory
9. Give language needing TM power
10. Give CFL that PDA accepts

”Why?” Questions Professors Love

1. Why DFA cannot recognize balanced parentheses?
2. Why stack increases computational power?
3. Why finite automata limited?
4. Why Turing machines model general computation?
5. Why undecidable problems exist?
6. Why pumping lemma works?
7. Why regular expressions correspond to automata?
8. Why parsing needs CFGs?
9. Why nondeterminism powerful conceptually?
10. Why computation has theoretical limits?

Computer Networks (CN)

Core Concepts

1. OSI layers and responsibilities. TCP/IP vs OSI.
2. TCP vs UDP - reliability, speed, use cases.
3. Three-way handshake: $\to$ -ACK $\to$ .
4. What is flow control? What is congestion control?
5. AIMD and slow start intuition.
6. What happens when you type a URL in a browser (DNS $\to$ TCP $\to$ HTTP)?
7. Hub vs switch vs router.
8. IPv4 vs IPv6 - why is IPv6 needed?
9. Why does packet switching scale better than circuit switching?

Database Management Systems (DBMS)

Core Concepts

1. Relational model: relation, tuple, attribute, schema vs instance.
2. Primary key vs candidate key vs foreign key.
3. WHERE vs HAVING. GROUP BY vs ORDER BY.
4. INNER JOIN vs OUTER JOIN vs self join.
5. DELETE vs DROP vs TRUNCATE.

Normalization

1. What is normalization? Why is it needed?
2. Functional dependency, partial dependency, transitive dependency.
3. Explain 1NF, 2NF, 3NF, BCNF. Difference between 3NF and BCNF.
4. Can denormalization improve performance?

Transactions & Indexing

1. ACID properties.
2. Dirty read, non-repeatable read, phantom read.
3. Why are B+ trees preferred over BSTs and hash tables for indexes?
4. Clustered vs non-clustered index.
5. Why can joins be computationally expensive?

Machine Learning & AI

ML Basics

1. Bias vs variance. Underfitting vs overfitting.
2. L1 vs L2 regularization.
3. Why does gradient descent work?
4. Precision vs recall.

Deep Learning

1. Backpropagation intuitively.
2. Why are activation functions needed? ReLU and dying ReLU.
3. Vanishing gradient problem.
4. Why does batch normalization help?
5. Softmax intuition. Why cross-entropy loss?

Transformers & NLP

1. What is the attention mechanism?
2. Why are transformers effective for NLP?
3. Why do transformers replace RNNs?
4. Curse of dimensionality.

Signals, Systems & Electronics

LTI Systems

1. What is an LTI system? (A system that satisfies both Linearity (superposition: $\alpha x_1+\beta x_2\to \alpha y_1+\beta y_2$) and Time Invariance (a time shift in input causes the same time shift in output).)
2. What makes a system linear?
3. What makes a system time-invariant?
4. Is $y(t)=tx(t)$ an LTI system? ( is linear (superposition holds), but not time-invariant: if $x(t)\to y(t)=tx(t)$, then $x(t-t_0)\to tx(t-t_0)$, but we expect $(t-t_0)x(t-t_0)$. The factor $t$ breaks time-invariance. So it is linear but not LTI.)
5. If two periodic signals are added, is the result always periodic? (Only if the ratio of their periods is rational.)

Fourier Series & Fourier Transform

1. What is the Fourier Series? (Representation of a periodic signal as a sum of sinusoids (harmonics). Exists for periodic, Dirichlet-condition-satisfying signals. Gives a discrete spectrum.)
2. What is the Fourier Transform? (Extension to aperiodic signals. Gives a continuous spectrum $X(f)={\int }_{-\infty }^{\infty }x(t)e^{-j2\pi ft}dt$.)
3. When to use which? (Fourier Series for periodic signals; Fourier Transform for aperiodic (or finite-energy) signals.)

Laplace Transform

1. Why does the Laplace Transform exist? Why is the Fourier Transform not enough?

• The Fourier Transform requires the signal to be absolutely integrable ($\int |x(t)|dt<\infty$), which many signals (e.g., growing exponentials, unstable system responses) are not.
• The Laplace Transform introduces a convergence factor $e^{-\sigma t}$ (Region of Convergence), allowing analysis of a broader class of signals including those that do not have a Fourier Transform.
• It also naturally encodes initial conditions, making it ideal for solving differential equations and analysing system stability via pole locations in the $s$-plane.

Sampling & Aliasing

1. What is sampling? (Converting a continuous-time signal to a discrete-time signal by measuring at regular intervals $T_s$.)
2. What does sampling do in the frequency domain? (Sampling at rate $f_s=1/T_s$ causes the spectrum to become periodic, repeating at multiples of $f_s$. Mathematically: $X_s(f)=f_s{\sum }_{k}X(f-k{f}_s)$.)
3. What is aliasing? (When $f_s<2f_{\max }$ (Nyquist criterion violated), the periodic spectral copies overlap, making it impossible to recover the original signal. High-frequency components appear as low-frequency “aliases.”)

Amplitude & Frequency Modulation

1. AM vs FM - which requires more bandwidth? (FM requires more bandwidth than AM for the same message signal. AM bandwidth $=2W$ (where $W$ is message bandwidth). FM bandwidth $\approx 2(W+\Delta f)$ by Carson’s rule, where $\Delta f$ is the frequency deviation - always $\ge 2W$.)
2. Draw an AM-modulated signal in the frequency domain. (An AM signal $[A_c+m(t)]\cos (2\pi f_{c}t)$ has: a carrier impulse at $\pm f_c$, and sidebands (copies of $M(f)$ shifted to $\pm f_c$). Total bandwidth $=2W$.)
3. Which is non-linear - AM or FM? (FM is non-linear (frequency is a non-linear function of the message). Standard AM (DSB-LC) is linear in the envelope sense, though generation can involve multiplication. Angle modulation (FM/PM) is inherently non-linear.)
4. Explain Carson’s rule. (FM bandwidth $\approx 2(\Delta f+W)$, where $\Delta f=k_f\cdot A_m$ is the peak frequency deviation and $W$ is the message bandwidth.)
5. Why is there an extra $+W$ (or $+B$) term in Carson’s rule? (Even at the extremes of frequency swing $\pm \Delta f$, the instantaneous frequency still occupies the message bandwidth $W$ around those extremes. The $+W$ accounts for the fact that the modulated signal isn’t an impulse at $f_c\pm \Delta f$; it spreads by $W$ on each side. Without it, you’d underestimate the actual spectral occupancy.)

MOSFET & Circuit Theory

1. Draw the Voltage Transfer Curve (VTC) for a MOSFET inverter. (The VTC plots $V_{out}$ vs $V_{in}$. Key regions: (1) $V_{in}<V_{th}$: MOSFET off, $V_{out}=V_{DD}$ (high). (2) $V_{in}$ near $V_{th}$: transition region (high gain, $V_{out}$ drops steeply). (3) $V_{in}>V_{th}$ (large): MOSFET in triode/saturation, $V_{out}\approx 0$ (low). Noise margins $N{M}_H$ and $N{M}_L$ are read from the unity-gain points.)
2. How does a capacitor behave in small-signal analysis? ( small-signal (AC) analysis, a capacitor is replaced by its impedance $Z_C=\frac{1}{j\omega C}$. At midband frequencies (where coupling/bypass capacitors are assumed large), capacitors are treated as short circuits (zero impedance). At high frequencies, internal device capacitances ($C_{gs}$, $C_{gd}$) are included as short circuits relative to their impedance.)
3. Explain Thévenin’s and Norton’s theorems.

• Thévenin: Any linear two-terminal network can be replaced by a voltage source $V_{th}$ (open-circuit voltage) in series with $R_{th}$ (resistance seen at terminals with independent sources killed).
• Norton: Equivalent is a current source $I_N=V_{th}/R_{th}$ in parallel with $R_{th}$.
• Useful for simplifying circuits and finding operating points for nonlinear elements.

4. What are oscillators? Draw an RC oscillator (RC phase-shift oscillator). (An oscillator is a circuit that generates a periodic output without an external AC input, by using positive feedback with unity loop gain (Barkhausen criterion: $|A\beta |=1$ and $\angle A\beta =0°$). An RC phase-shift oscillator uses 3 RC sections each providing 60° phase shift (total 180°) plus the 180° from the inverting amplifier, achieving 360° total. Frequency of oscillation: $f=\frac{1}{2\pi RC\sqrt{6}}$.)

Meta / Thinking-Oriented Questions

1. Can you reason from definitions?
2. Can you recover after mistakes?
3. Can you think aloud?
4. Can you derive intuitively instead of memorising?
5. Can you provide counterexamples?
6. What breaks if assumptions are removed?
7. Can you defend your answer under follow-ups?

Common Trap Questions (All Topics)

1. Can basis be dependent?
2. Can DFS use a queue?
3. Can an onto function exist if codomain is larger than domain?
4. Can Dijkstra work with a negative cycle?
5. Can CDF exceed 1?
6. Is every spanning set a basis?
7. Does zero correlation imply independence?
8. Can variance be negative?
9. Why can XOR alone not form all logic gates?
10. Why does greedy fail sometimes?
11. Is multithreading always faster?
12. Can a page fault occur even if memory is available?
13. Can deadlock occur with only one thread?
14. Can starvation happen without deadlock?
15. If RAM were infinite, would OS concepts change?
16. Can a compiler be written in the language it compiles?

Puzzle / Thinking Questions

1. How would you measure half a cylinder of water without tools?
2. How many tennis balls fit in this room?
3. Find a duplicate number in an array with minimum memory.
4. You have 8 balls, one heavier - minimum weighings needed?
5. How would you efficiently test whether a remote server is alive?

Key Takeaways for Future Aspirants

2026 Batch Advice to You

1. Linear Algebra is the most common pitfall. Almost every ECE panel asked LA. Know matrices, eigenvalues, eigenvectors, and vector spaces deeply - not just definitions.

2. Don’t panic if you can’t answer immediately. The panel gives hints. Engage with the hint, reason out loud, and try. Composure under pressure is itself being evaluated.

3. Approach $>$ Answer. Thinking out loud and demonstrating a coherent reasoning process impresses more than a memorized correct answer. Multiple candidates who fumbled final answers still got through.

4. Give your intro proactively. Many panels start with questions directly. If they don’t ask for an introduction, bring up your highlights yourself - research, projects, lab interests.

5. Mention IIITH lab interests explicitly. Even if not asked “Why IIIT-H?”, steer the conversation to name 1-2 labs you want to join (CVIT, PreCog, CSTAR, RRC, etc.) and why. At least one candidate regretted not doing this.

6. Research papers are a strong differentiator. Interviewers responded visibly to publications. If you have any, name the venue clearly - they may recognise it and it can enter their final deliberation.

7. Duration is not a signal. Interviews in 2026 ranged from 4 minutes to 30+ minutes. Both short and long interviews resulted in selections.

8. Be strategic about which subjects you list. Panelists tend to ask from what you declare. Do not volunteer subjects you have not studied deeply.

9. Do not rely entirely on transcripts. Questions change significantly year to year and panel to panel. Understand concepts; do not just memorise past answers.

10. Do mock interviews. Multiple candidates credited the mock with preventing far worse performance. It conditions you to the format and reduces freeze-ups.

11. Case studies can replace direct questions. At least one candidate received a probability + networks case study as the basis for the entire interview. Be ready to reason from a scenario, not just recite theory.

12. If the panel asks if you have questions for them, have one ready. Two candidates - one who froze when asked, one who simply said no - both regretted it. Prepare one genuine question about research at IIITH.

IIIT-H Lab & Research Info (from 2026 seniors)

Lab / Group	Notes
PreCog	CV, NLP, MDL, and other AI branches. Most competitive: 90+ applicants for $\sim$5 spots. Gives tasks before interview.
CVIT	Computer Vision and Intelligence. Good fit for robotics + AI interests.
RRC	Robotics Research Center.
CSTAR	Mentioned by multiple aspirants as a target.
AlphaGrep	Quant research group (not a lab). Formed by Prof PK and Prof Bodas. Focus: quantitative finance + applied math. Only Bodas-track students go here.

• Lab selection happens mid-2nd year via a form. You pitch a research idea to an advisor; they agree or decline.
• Research and thesis run from summer after 2nd year through to thesis defence (typically 5th year).
• Advisors generally will not let you research too far outside their own domain.
• Approach labs early (even day one of joining) - initiative is noticed and valued.
• You can join PreCog externally via Saral AI as well.

—

Historical LEEE Dates (2016-2026)

Note on Result Delays

Results can be delayed when UGEE interview dates clash with the expected LEEE result announcement window.

Dates are accurate to the best of available knowledge. Earlier years (2014-2016) in particular may contain inaccuracies.

Year	Written Test	Test Results	Interview	Final Results
2014	13 Apr	17 May	-	24 May
2015	19 Apr	24 May	-	29 Jun
2016	17 Apr	14 May	-	27 Jun
2017	30 Apr	24 May	9 Jun	12 Jun
2018	21 Apr	30 Apr	-	31 May
2019	28 Apr	9 May	25 May	23 Jun
2020	24 Jun	6 Jul	17 Jul	21 Jul
2021	3 Jul	10 Jul	20 Jul	28 Jul
2022	16 Apr	24 Apr	13 May	16 May
2023	6 May	17 May	27 May	6 Jun
2024	4 May	18 May	25 May	3 Jun
2025	19 Apr	12 May	24 May	4 Jun
2026	2 May	11 May	23 May	8 Jun

Selection Timeline (2026)

Date / Period	Event
May 23-24, 2026	Interviews conducted at IIITH
Early June 2026	Results expected (based on previous year pattern of 8-10 days)
Post-result	Emails sent to all candidates - accepted and rejected

Related

• previous-year-questions-pyqs
• leee-2026-memory-based-question-bank