DSA mock interviews for the part you can't practice alone.
Clarify the prompt, choose a direction, say the quiet parts out loud, and defend your trade-offs — against a voice-based AI interviewer that pushes back.
Merge Intervals
Merge overlapping intervals and return the non-overlapping result.
interviewer
Walk me through the brute-force approach first.
you
I would compare every pair, then sort to avoid repeating that work.
interviewer
What makes sorting the better starting point?
01 def merge(intervals):
02 intervals.sort()
03 merged = []
04 for interval in intervals:
05 if not merged or ...:
06 merged.append(interval)follow-up
What does sorting buy you?
“Don't just give me the answer. Tell me what you ruled out.”
§01
Definition
What MockDSA is
MockDSA is a voice-based AI mock interview tool for data structures and algorithms interviews. You speak your reasoning out loud while writing code in a browser editor, an AI interviewer asks follow-up questions and pushes back in real time, and when the session ends you get a report scoring you from 0 to 5 on 8 criteria, with an overall hire verdict and specific instructions on what to fix.
- Session length
- 30, 45 or 60 minutes
- Problem bank
- 75 Blind 75, or your own
- Languages
- 13, including Python, Java, Go
- Scored on
- 8 criteria, 0–5
§02
The gap
LeetCode gives you a problem. The screen gives you a conversation.
A correct solution can still arrive wrapped in a confusing explanation, an unexplored edge case, or ten minutes of silence. Those are the reps most people skip because there's no easy way to do them by yourself.
MockDSA gives the conversation somewhere to go. It listens to your reasoning, asks the next question, and keeps the moment just uncomfortable enough to be useful.
Make the problem smaller.
Clarify inputs, outputs, constraints, and the examples you actually need.
Keep the plan visible.
Explain the trade-off before the interviewer has to guess what you meant.
Recover without disappearing.
Respond when a follow-up changes the shape of the problem.
§03
Session
Thirty minutes, narrated.
- -00:30
Before the clock starts
Paste a problem of your own or pull one from the 75 in the bank, choose 30, 45 or 60 minutes, and pick a voice.
- 00:00
The problem lands
The interviewer reads you in and then stops talking. What happens next is on you — there's no “click here to begin” scaffolding.
- 01:30
You ask, or you don't
Clarifying questions are scored, and so is skipping them. Most people discover here that they normally start coding far too early.
- 04:00
You commit out loud
State an approach and why. The editor and scratchpad are open the whole time, and both are read later as evidence.
- 12:00
The follow-ups start
Why sorting. What breaks on an empty input. What it costs. It will nudge, but it will not hand you the key data structure.
- 30:00
The clock stops
Everything you said and typed goes to an evaluator that scores 8 criteria out of 5 and returns a verdict.
§04
Coverage
What you can practice
The built-in bank is the Blind 75 — 75 problems grouped into the 10 shapes that keep reappearing in coding screens. Or paste your own, which is what most people do once they have a specific company in mind.
- 01Arrays
- 02Strings
- 03Linked Lists
- 04Trees
- 05Graphs
- 06Dynamic Programming
- 07Intervals
- 08Matrices
- 09Heaps
- 10Binary & bit manipulation
$ languages — Python · JavaScript · TypeScript · Java · C++ · C · C# · Go · Rust · Ruby · PHP · Swift · Kotlin
§05
Report
When it's over, you should know what to do next.
The result is not a gold star for showing up. It breaks down how you performed across the process: the conversation, the code, the scratchpad, and how much guidance you needed.
MockDSA evaluates the same things an interviewer can observe while you work — not just the final answer.
EXAMPLE REPORT
Merge Intervals
Hire
3.9 / 5
4 OF 8 CRITERIA SHOWN
NEXT REP
State the brute-force option before optimizing. It makes the trade-off easier to follow — and gives the interviewer something concrete to challenge.
§06
Scoring
How MockDSA scores you
Every session is scored from 0 to 5 on each of these 8 criteria. The evaluator reads the whole timeline — what you said, what you typed, and when — so a correct final answer does not cancel out fifteen minutes of confusion getting there.
- 01
Problem Solving
__ / 5Whether you found a workable approach yourself, and how much of the path you reasoned out versus stumbled onto.
- 02
Communication & Thinking Out Loud
__ / 5Whether an interviewer could follow your reasoning as it happened, without having to ask what you meant.
- 03
Code Quality
__ / 5Whether the implementation is readable and structured, or a sequence of unexplained rewrites.
- 04
Planning & Approach
__ / 5Whether you settled on a direction before coding and said why, rather than trial-and-error without a hypothesis.
- 05
Time & Space Complexity Analysis
__ / 5Whether you can state the cost of your solution and justify it, not just recite a Big-O label.
- 06
Use of Scratchpad/Notes
__ / 5Whether you used the scratchpad to hold state — examples, invariants, edge cases — when the problem needed it.
- 07
Handling Hints & Guidance
__ / 5How much help you needed, and whether you took a nudge and carried on independently or waited to be led.
- 08
Solution Correctness
__ / 5Whether the final solution actually works, including the edge cases you were expected to find.
§07
Verdicts
What the verdicts mean
The criteria roll up into one of five verdicts, using the vocabulary a real debrief uses. Drawn to scale, the bands are lopsided on purpose: half the axis is No Hire, and the entire difference between “arguable” and “obvious yes” is the last point and a half.
- Strong Hire
- 4.5+
- Solved the problem optimally, with communication that carried the room.
- Hire
- 3.5 – 4.5
- Solved the problem and explained it well enough to be easy to follow.
- Lean Hire
- 3.0 – 3.5
- Mostly solved it. A real committee would still be arguing about you.
- Lean No Hire
- 2.5 – 3.0
- Partial solution, or communication that got in the way of the work.
- No Hire
- Below 2.5
- Did not solve it, needed the answer supplied, or a red flag in the process.
§08
Scope
It's deliberately not LeetCode.
MockDSA assumes you've studied. It exists for the uncomfortable part after the studying — which means there are things it pointedly will not do.
- No execution
- The focus is your reasoning, not a green test-run badge.
- No hand-holding
- You get follow-ups and pushback, not a tutorial.
- No pretending
- It's AI, not a human interviewer — and we say so.
§09
Cost
What it costs
One credit is one interview. The interview is identical on every plan — you are only choosing how many reps you want in a month, so the honest way to compare them is per interview.
- Small
- $19/mo
- 5 interviews
- $3.80 per interview
- Medium
- $39/mo
- 12 interviews
- $3.25 per interview
- Large
- $69/mo
- 24 interviews
- $2.88 per interview
One-off credit packs are available without a subscription. See full pricing.
§10
Questions
A few practical questions
What is MockDSA?
MockDSA is a voice-based AI mock interview tool for data structures and algorithms interviews. You talk through a coding problem out loud while writing code in a browser editor, the interviewer asks follow-up questions as you go, and afterwards you get a report scoring you on 8 criteria with an overall hire verdict.
How long is a mock interview?
You choose 30, 45 or 60 minutes when you start a session. The default is 30 minutes, which matches a typical technical phone screen.
Do I have to use voice?
Yes. The point is to practice communicating your reasoning in real time, so a microphone is required. Interviews are currently conducted in English.
Can I use my own problem?
Yes. You can paste in any problem — from LeetCode, Blind, a recruiter email, or anywhere else — or pick from the built-in bank of 75 Blind 75 problems.
Which programming languages are supported?
The editor supports 13 languages: Python, JavaScript, TypeScript, Java, C++, C, C#, Go, Rust, Ruby, PHP, Swift, Kotlin. Use whichever one you would use in a real interview.
How is the interview scored?
Every session is scored from 0 to 5 on 8 criteria: Problem Solving, Communication & Thinking Out Loud, Code Quality, Planning & Approach, Time & Space Complexity Analysis, Use of Scratchpad/Notes, Handling Hints & Guidance, Solution Correctness. Those roll up into an overall score and one of five verdicts, from Strong Hire down to No Hire.
What does a verdict like “Lean Hire” mean?
The verdicts mirror the language real interview debriefs use. Strong Hire is a 4.5+ average, Hire is 3.5 to 4.5, Lean Hire is 3.0 to 3.5, Lean No Hire is 2.5 to 3.0, and No Hire is below 2.5. Lean Hire means you mostly solved the problem and communicated decently, but a real committee would still be arguing about you.
Does MockDSA run or test my code?
No. There is no execution and no test-runner badge. Correctness is judged by the interviewer from your code and your reasoning, which is how a real interview works too.
Does it cover system design or behavioural interviews?
No. MockDSA is deliberately scoped to data structures and algorithms interviews only.
Do I need to install anything?
No. Everything runs in the browser — the interviewer, the code editor, and the scratchpad. You only need a microphone.
What does MockDSA cost?
Plans are monthly and credit-based, where one credit is one interview: Small at $19 for 5 interviews a month, Medium at $39 for 12 interviews a month, Large at $69 for 24 interviews a month. One-off credit packs are also available without a subscription.
What happens to my code and voice?
Session content is processed in real time and discarded when the session ends. Your scored report is saved to your dashboard. See the homepage FAQ for provider details.
Last reviewed .