Skip to content

Mock interviews: where, how, how many

For anyone who can solve problems alone but has not practiced solving them in front of someone. You get free and paid options with prices, a run sheet and scoring rubric for peer mocks, and a schedule for how many to do and when.

Jugal's reason: "Knowing the answer and being able to say it calmly under pressure are two different things, and you only close that gap with reps" (Want a Job in the Next 30 Days?). He makes the same point in The job-search tool stack I'd actually use in 2026.

Why mocks change the result

Finding Source
Candidates were "almost 2X as likely to pass" Facebook after at least five practice interviews interviewing.io, practice gap
Predicted phone-screen pass rates with 5+ prior real technical interviews vs 1 to 4: Amazon 81% vs 65%, Facebook 71% vs 40%. Google: 76% with 3 interviewing.io mocks vs 69% with 2 interviewing.io, readiness survey
Among students who practiced regularly, "there was no difference in performance between elite schools and non-elite schools" interviewing.io, practice gap
Only about 20% of candidates perform consistently from interview to interview Same post
Candidates averaging about 3.0 out of 4 still fail single interviews about 22% of the time interviewing.io, performance is noisy
Self-ratings vs interviewer ratings: R-squared only 0.24. You cannot tell how you did interviewing.io, self-assessment
Women quit interview practice 7 times more often than men after one bad interview interviewing.io, practice gap

What to do with it:

  1. Do at least 5 mocks before a loop that matters.
  2. Get written feedback every time. Your own memory of the session is unreliable.
  3. Never judge yourself by one bad session. Book the next one the same day.
  4. Real interviews count as reps too. Schedule lower-priority companies first, so your top choice is not your first real interview.

Options, free first (as of Oct 2026)

Option Type Price Best for How to use it
Aced Practice (formerly Pramp) Peer, matched for you Free monthly credits; paid membership for unlimited (plans) Weekly peer reps with strangers Book a DSA session. You and your partner interview each other for about 30 minutes each
A friend, classmate, or Discord partner Peer, self-run Free Regular reps on your schedule Run the 60-minute format below. Find partners on cscareers.dev Discord (164K members), CS Majors (31.8K), NeetCode (46K), or the smaller Tech Interview Handbook Discord (2.8K)
Alumni or engineers at your target company Peer, expert Free (ask politely) Company-specific calibration Send the partner request below. Offer to return the favor, and keep it to 45 minutes
interviewing.io AI Interviewer AI Free Learning the format before your first human mock "It's free, it simulates a FAANG-style interview, and it can do both coding and system design" (FAQ). Do your first 2 to 3 mocks here
CodePath Technical Interview Prep Course with mocks "No-cost to CS students" (check eligibility on the page) Structured 10-week program, 5 to 10 hours a week Apply a term before your recruiting season. It teaches UMPIRE, and the intermediate course includes mock interviews
CoderPad sandbox Environment only Free Using the same editor as many real interviews Share it with your mock partner. Practice without running code to mimic Google, Meta, and Amazon
Free Mock Interview AI, voice Free, no login Saying answers out loud, mostly behavioral Pick your target role and answer out loud. Read the report for where you ramble
Your university career center Human Free for students Behavioral rounds, nerves Book one early in the season
interviewing.io mocks Human, FAANG engineers Paid: "Interviews start at $179"; full refund if unhappy (FAQ) Calibration right before a big loop Anonymous, voice only. Book 1 or 2 in the final 2 weeks; pick an interviewer from your target company. Jugal used it before Amazon: explaining out loud "helped me articulate my thought process clearly during actual Amazon rounds" (post)
Meetapro Human marketplace Paid; prices vary by interviewer (filter by price) A specific company or round type Filter by company and round, read reviews, book 1 before an onsite
Aced coaching (formerly Exponent) Human coaches Paid session packs A final pre-onsite mock with a senior coach Book one session only, in the last 2 weeks
Preplaced Mentorship marketplace (India) Paid, free trial Long-term 1:1 mentorship from 600+ mentors at companies like Google, Amazon and Meta Use the free trial to test a mentor before you pay. Ask for mock interviews in the plan
Hello Interview Premium Guided practice, no humans $47 per month, $79 per year, $279 lifetime (sale prices as of Oct 2026; list $59, $99, $349) AI-enabled coding and system design practice Its human mocks ended May 31, 2026 (notice). Buy only if you also need system design
LeetCode Premium Timed assessments $35 per month or $159 per year (sale prices as of Oct 2026) OA rehearsal with company-style timed sets Buy one month right before your loop. Use the mock assessments

Watch out: Final Round AI appears in some older mock-interview lists, including two of Jugal's posts. Its main product, "Interview CoPilot", "listens to your live interview and delivers real-time answers in a private window only you can see", with a "Stealth Mode". Using that in a real interview breaks Google, Meta, Amazon, Microsoft and Anthropic rules, so use AI for practice only.

Changed or gone (do not follow older links to these):

  • Pramp now runs on Aced Practice.
  • Exponent became Aced in August 2026 (announcement).
  • Google Interview Warmup is retired.
  • interviewing.io no longer offers free human peer mocks.

Which option to use at each stage

Stage Use Why
First 2 weeks of prep AI Interviewer, Free Mock Interview Learn the format with no social cost
Week 3 onward Aced Practice or a Discord partner, 1 to 2 a week Habits form through reps with a real person
Final 2 to 3 weeks before a loop Add 1 to 2 experienced interviewers (paid or alumni) Calibration from someone who has run real loops
Final week 2 to 3 peer mocks in the exact format of your loop Rehearsal, timing, setup
AI-enabled round coming Hello Interview sandbox or the company's practice link You need the real tool, not a whiteboard

How to run a peer mock

Before the session

  1. Agree on the format: language, length, and which run sheet below.
  2. Each person picks 2 to 3 problems the other has not seen, from different topics, and knows the different approaches to each (TIH interviewer cheatsheet).
  3. Avoid famous problems word for word: change a constraint or wrap it in a story. Pull problems from Problem lists or a company page. Custom questions test real understanding: in an interviewing.io experiment, candidates secretly using ChatGPT passed 73% of verbatim LeetCode questions but only 25% of custom ones (experiment).
  4. Set up a video call, a shared CoderPad sandbox (agree not to run the code), a visible timer, and the scoring sheet below.
  5. The interviewer reads the TIH interviewer cheatsheet once.

Run sheet A: one candidate, 60 minutes (most realistic)

Clock What happens
0:00 to 0:02 Interviewer explains format and language, asks "Have you seen this question before?"
0:02 to 0:45 The interview. Candidate follows the 45-minute framework
0:45 to 0:50 Interviewer scores silently on the sheet
0:50 to 1:00 Feedback: 1 strength, 2 fixes with timestamps. Copy the code out before closing

Swap roles next session.

Run sheet B: swap in one session, about 75 minutes (the Aced format plus scoring)

Clock What happens
0:00 to 0:02 Setup, start timer
0:02 to 0:30 Round 1: A interviews B
0:30 to 0:33 A scores silently
0:33 to 1:01 Round 2: B interviews A
1:01 to 1:04 B scores silently
1:04 to 1:15 Exchange feedback both ways from the sheets

Interviewer script

OPENING
"We have 45 minutes. One problem, maybe a follow-up. Code in [language]. Code won't run, so trace by hand.
 Leave 5 minutes at the end for questions. Have you seen this question before?"

GIVING THE PROBLEM
Read the problem with one small example. Do not volunteer constraints. Answer only what they ask.

WHILE THEY WORK
Stay quiet unless asked. Note timestamps: clarify ended, coding started, coding ended, testing ended.
Do not point out bugs. Ask: "What happens if the input is [X]?"
If they say "done" without testing: "How would you check this works?"

AFTER THE CODE
"What's the time and space complexity?"
"What if [constraint changes]?"

CLOSE
"Any questions for me?" Then stop the timer and score before you talk.

Hint ladder

Give the smallest hint that unblocks them, and log the level and minute. Wait for 3 to 5 minutes with no progress before the first hint (a rule of thumb for this site). The TIH cheatsheet says that if the candidate "is still stuck after providing hints, provide the solution and move to coding so that you can get coding signals."

Level Hint type Example (Subarray Sum Equals K)
1 A question about their own work "Where does your brute force repeat work?"
2 A direction "Could you reuse sums you already computed?"
3 A technique "Think about prefix sums and a hash map."
4 The approach "Count earlier prefixes equal to the current prefix minus k." Then watch them code

Level 1 costs little on the rubric. Level 4 means the problem-solving score is a 2 at best.

Scoring rubric

Four dimensions, 1 to 4 each, matching the Tech Interview Handbook rubric and Meta's four areas. The anchors below were written for this site from the TIH signals. Use the same anchors every time so scores compare across weeks.

Score Problem solving Coding Verification Communication
4 Strong hire Clarified fully, gave brute force and optimal with correct complexity, no hints, time for follow-ups Working code matching the plan, clean names, helpers, idiomatic, no syntax errors Traced unprompted, covered edge cases, found and fixed own bugs Talked the whole way, clear and short, took feedback well
3 Hire Reached optimal with one small hint; complexity correct Working code with minor slips, readable Traced at least one example and the main edge cases Explained approach before coding, rare silences
2 No hire Needed a technique-level hint, or settled on brute force only Bugs left, messy structure, or partial pseudocode Tested only when prompted, missed obvious edge cases Long silences or half-thoughts, coded before explaining
1 Strong no hire Could not reach a working approach even with the approach given Code does not match the plan or does not compile No testing Unresponsive, argued with hints

Copy this sheet into a doc for every mock:

Candidate: [name]   Interviewer: [name]   Date: [date]   Question: [title]   Seen before? [Y/N]
Scale: 4 strong hire, 3 hire, 2 no hire, 1 strong no hire

PROBLEM SOLVING        [1-4]
  [ ] Asked clarifying questions before coding
  [ ] Gave brute force with complexity
  [ ] Reached an optimized approach
  [ ] Stated time and space correctly
  [ ] Hints used: level [0-4] at minute [mm]

CODING                 [1-4]
  [ ] Code matches the discussed approach
  [ ] Syntactically correct, no pseudocode
  [ ] Clear names, helper functions, no globals
  [ ] Handles edge cases explicitly
  [ ] Uses language idioms appropriately

VERIFICATION           [1-4]
  [ ] Traced an example line by line without being asked
  [ ] Listed and checked edge cases
  [ ] Found and fixed own bugs
  [ ] Did not say "done" before testing

COMMUNICATION          [1-4]
  [ ] Talked through the approach before coding
  [ ] Narrated while coding (no silence over 2 minutes)
  [ ] Finished thoughts aloud (no half-thoughts)
  [ ] Took hints well, did not argue

Timeline: clarify ended [mm:ss]  coding started [mm:ss]  coding ended [mm:ss]  testing ended [mm:ss]
Top strength: [one line]
Fix 1: [one line, with timestamp]
Fix 2: [one line, with timestamp]
Overall: [strong hire / hire / no hire / strong no hire]

Feedback rules

  1. Score before you speak. Your first impression drifts once you start talking.
  2. Name behavior, not personality: "Silent from 12:10 to 15:30" beats "seemed nervous".
  3. Give one strength and two fixes. More than two fixes will not stick.
  4. Write it down and send it. Self-assessment is unreliable (R-squared 0.24), so the candidate needs your notes.
  5. Tie each fix to the code quality checklist or the framework step it belongs to.

Partner request template

Hi [name], I'm prepping for [company] [intern / new grad] interviews on [date]. Looking for a weekly mock partner:
60 minutes, one of us interviews per session, CoderPad, scored with the Tech Interview Handbook rubric.
I'm free [days and times, time zone]. I'm working through [NeetCode 150 / Grind 75]. Interested?

Practice with a chat model as the interviewer

Use this when no partner is free. Never use any outside AI during a real interview unless the company allows it.

Act as a [Company] software engineer running a 45-minute coding interview for a [intern / new grad] role.
Rules:
- Give me one [Medium] problem in a short story wrapper, not a famous LeetCode title. Do not show a solution.
- Answer my clarifying questions briefly, like a real interviewer. Do not volunteer constraints.
- If I make no progress for 3 of my messages, give the smallest possible hint.
- When I say "done", ask me to test my code. Then ask for time and space complexity and one follow-up.
- At the end, score me 1 to 4 on problem solving, coding, verification and communication.
  Give one strength and two fixes, quoting my own words.

Mirror the real format

Target Mock format Source
Google 45 minutes, one problem plus follow-ups, plain doc, no code execution. Add one whiteboard mock if your loop is onsite interviewing.io Google guide, Google page
Meta screen and classic rounds Two problems in about 35 to 40 minutes, CoderPad, no code execution interviewing.io Meta guide, Meta page
Meta AI-enabled round 60 minutes, multi-file starter code, AI chat allowed, tests run Hello Interview
Amazon 10 to 15 minutes of Leadership Principle questions, then one coding problem. Mix in a design-a-class prompt for the logical-and-maintainable round Amazon page, LPs
Microsoft 45 minutes; code may run; interviewer asks about testing and error conditions Microsoft technical interviewing
Online assessment Solo, timed, on the OA platform's practice test OA strategy
Behavioral 45 minutes of STAR stories with follow-ups Behavioral, story bank
System design (if your level gets it) 45 to 60 minutes on a whiteboard tool System design framework

How many mocks and when

When How many Type Goal
Weeks 1 to 2 of active prep 1 per week AI Interviewer Learn the format
Week 3 until 3 weeks before the loop 1 to 2 per week Peer, alternating interviewer and candidate Build the habits on the framework page
Final 2 to 3 weeks At least 5 in total with people who are not close friends, including 1 to 2 experienced interviewers Peer plus paid or alumni Calibration on unseen problems
Final week 2 to 3 Exact format of your loop Timing and setup
Day before 0 Light review of your phrase bank and story bank Rest

The minimum of 5 comes from the interviewing.io data above: "You need at least five interviews to move the needle, and they need to happen systematically and on a schedule." Jugal's own plans fit inside this: 3 timed 45-minute mocks in week 5 of each 5-week company plan (Meta, Amazon, Google, Apple, Netflix; all five in one post: Optimizing your interviews at FAANG), and two sessions a day on days 3 to 7 of week 5 in Your 6-Week Amazon Interview Roadmap.

Ready signal

You are ready when 3 mocks in a row, on problems you have not seen, score 3 or higher on all four dimensions. This is a rule of thumb for this site, not a published standard. Strong candidates still fail about 22% of single interviews, so judge over several mocks, never one.

Not ready? Postpone

Recruiters usually agree. Aline Lerner of interviewing.io: "If you need two months, ask for two months" (post). Her script, verbatim:

I'm really excited about interviewing at [company name]. Unfortunately, if I'm honest, I haven't had a chance to practice as much as I'd like. I know how hard and competitive these interviews are, and I want to put my best foot forward. I think I'll realistically need a couple of months to prepare. How about we schedule my interview for [date]?

After every mock

  1. Save the code, the timeline, and the scores in your mock log.
  2. Answer the self-review questions below in writing, adapted from Jugal's Amazon roadmap.
  3. Listen to the recording. Jugal: "Notice how you say 'um' forty times? Fix it. Notice how you went silent for 3 minutes? Never do that again."
  4. Re-solve any problem that took over 30 minutes, in 25 minutes or less (Amazon prep post). Do it 3 days later, the gap Jugal uses in his Amazon roadmap.
  5. Record a 2-minute audio explanation of that solution, then cut it until it is clear. Every week 5 in Jugal's company plans includes a recorded explanation. His Netflix plan caps it at 2 minutes, and his Google plan makes it a "proof sketch" of the time and space trade-offs.
  6. Pick one fix for the next mock and write it at the top of your next scoring sheet.
SELF-REVIEW (answer in writing within 30 minutes)
1. Did I talk enough?
2. Was my solution clear before I coded?
3. Did I sound like I knew what I was doing, or was I guessing?
4. Would you hire me based on this interview?   (ask the interviewer)
5. Where did I go silent, and for how long?
6. Which part took longer than its time box on the framework page?

MOCK LOG (one row per mock; keep it in your practice tracker)
Date | Partner / tool | Problem | Seen before? | PS | Code | Verify | Comm | Hint level | Top fix | Re-solved by (date)
[YYYY-MM-DD] | [name] | [title] | [Y/N] | [1-4] | [1-4] | [1-4] | [1-4] | [0-4] | [one line] | [date]

Keep it next to your problem log from How to practice.

Resources

Next: Companies