Mock interview
Mock Interview Drills That Get Data Freshers Hired
Stop rehearsing answers and start fixing weak spots. A practice-loop approach to mock interviews, with feedback checklists you can actually use.
Most mock interviews go nowhere because the feedback is too vague to use. You answer ten questions, someone says “good, but be more confident”, and next week you repeat the whole thing. What exactly are you supposed to fix on Tuesday afternoon?
Treat practice like debugging instead. Pick a small part of the interview, run it, write down what broke and change one thing. Then run it again. You’re not memorising a perfect speech. You’re finding the answers that fail under pressure and repairing them one at a time.
Pick one weak spot per session
A full mock interview is a poor practice unit when you’re starting out. Too many things can go wrong at once. The feedback becomes “speak louder, know your projects better, brush up SQL, and smile more”, which gives you four homework assignments and no useful place to start.
Pick the problem that would hurt you most in a real interview. These come up constantly in our practice rooms:
- You freeze when asked to walk through your project.
- You can write a
GROUP BYquery but go blank when asked why you used it. - You speak for 90 seconds when the question needed a 20-second answer.
- You can’t explain how a pandas
mergediffers from aconcat.
Notice how specific those are. “Bad at interviews” isn’t a drill. “My project answer falls apart after the first sentence” is.
Stay with that one weak spot for the week. If your project explanation is messy, don’t also try to repair your SQL narration and body language in the same session. You’ll end up with another round of vague feedback, except this time you’ll have supplied it yourself.
Run the drill in short loops, not long sessions
Twenty minutes, three times, beats one two-hour session. Long sessions feel productive because you’re tired afterwards. That isn’t evidence that your answers improved.
Use this loop.
1. Record yourself. Use your phone camera or laptop webcam and answer three questions about your chosen weak spot. No retakes. A retake hides the exact pause or ramble you came to fix.
2. Score the recording with a checklist, not a feeling. Watching yourself can be painfully awkward, so give your eyes and ears a job. For an explanation question, use this sheet:
[ ] Answer started in the first 15 seconds?
[ ] Named the actual tool, function or step?
[ ] Explained WHY, not just what?
[ ] Any filler over 3 seconds ("umm", "basically", "as such")?
[ ] Stopped when done, or kept rambling?
Five boxes. Tick or cross each one. Don’t replace the sheet with “that felt okay”. Your memory of an answer is usually much kinder than the recording.
3. Fix exactly one box. If you rambled, put a hard 45-second cap on the next three answers. If you never explained your reasoning, each answer must include a sentence beginning with “because”.
That second rule is deliberately hard to dodge. “We used merge” describes an action. “We used merge because both tables shared a customer ID and we needed matching rows in one table” gives the interviewer something to assess.
4. Record again the next day. Compare the recordings on that one box only. If you fixed the rambling but used more filler, that filler goes into the next round. Don’t declare the whole session a failure.
The gap matters. On the second recording, you’ll often hear the mistake just before you make it. That’s when you get a chance to stop and choose a better answer. Repeating the same response five times in one sitting mostly trains your memory, and a memorised answer tends to collapse as soon as the interviewer changes the question slightly.
The project drill: the one everyone avoids
Nine out of ten times, the weakest answer in a fresher mock is the project explanation. You built a dashboard or scraping script for a college submission, but the whole thing comes out as “we used Python and machine learning”. That sentence could describe almost anything.
Answer the same project question in three lengths. The content should change each time. Don’t deliver the same speech more slowly.
- 30 seconds: the problem, what you built and one result.
- 90 seconds: add one technical decision and explain why you made it.
- 3 minutes: add what broke and what you’d do differently now.
The 3-minute version is where interviewers find useful follow-up questions. If you run out of material after a minute, that’s information. You don’t know your own project well enough yet.
Go read your own code. Find a decision you made, even a small one. Why that input format? Why that table structure? What happened when the data was missing or messy? If you can’t answer those questions, another hour of speaking practice won’t rescue the project round.
The SQL and coding drill: talk while you type
In a data interview, someone may watch you solve a query problem. The query has to run, but the interviewer also needs to follow your thinking. Silence makes that impossible.
Practise speaking while you work, even when nobody else is in the room:
-- "First, checking the grain: one row per order or per item?"
SELECT order_id, COUNT(*) FROM order_items GROUP BY order_id;
-- "Multiple rows per order, so revenue needs SUM over items."
SELECT order_id, SUM(price * qty) AS revenue
FROM order_items GROUP BY order_id;
The first query checks what one row represents before the revenue calculation begins. That’s a useful habit because the table can contain several item rows for the same order_id.
What hurts candidates here is often the silence, not wrong SQL. Thirty seconds of typing without a word leaves the interviewer guessing whether you’re stuck or thinking. Say “checking the join keys” and your reasoning becomes visible.
One beginner habit needs special attention: jumping to JOIN before understanding the grain. Ask “what does one row represent?” every single time. Make it a reflex, not something you remember after getting a suspiciously large revenue total.
The feedback checklist for whoever is mock-interviewing you
Your friend or senior will probably default to encouragement unless you give them something concrete to check. Encouragement is pleasant. It won’t tell you why your answer wandered away from the question.
Share this before the session starts:
Answer quality
[ ] Answered the question asked, not a nearby one
[ ] Named concrete tools or steps, no vague "we used technology"
[ ] Gave a reason for at least one choice
[ ] Admitted a gap instead of bluffing
Delivery
[ ] Started within a few seconds of the question ending
[ ] No filler longer than 3 seconds
[ ] Stopped at the end instead of trailing off
Follow-up drill
[ ] Asked one follow-up and pushed for depth
That last line deserves attention. A mock without pushback doesn’t prepare you for “okay, but why not the other approach?”
Tell your mock interviewer to choose one claim from your answer and question it. If you say a tool was faster, ask faster compared with what. If you say a design was simpler, ask simpler for whom. You’re practising how to explain a choice after your first sentence has already been challenged.
Keep a failure log, not a confidence diary
After each drill, write one line containing the weak spot, the symptom, the fix and the retest date:
project explanation | rambles past 2 minutes | 45-second cap | retest Thursday
Four columns, five rows, one notebook page. That’s enough.
Students skip this because it feels like homework. It’s also the only part that gives you evidence across sessions. When you repeat the drill two weeks later and that line stays clean, you know the fix worked. That beats trying to reconstruct your progress from how confident somebody said you looked.
Small things that cost marks and take a minute to fix
Do these early. They take little practice compared with rebuilding a weak project answer.
- Say the full form once, then use the short form: “extract, transform, load. So the ETL step...”
- When you don’t know something, say what you’d check: “Haven’t used window functions much. Would look up
ROW_NUMBERand test it on a small table.” That’s a real answer. Bluffing your way into syntax you don’t know is not. - Keep your project open in another tab. “Let me show you” beats describing a screenshot from memory.
- Ask one question at the end about something you genuinely want to know, such as the team’s data or how work gets reviewed.
And no, a mock interview isn’t a performance you pass or fail. It’s a measurement. A bad measurement tells you where to look next.
If you want structured practice with someone running these loops alongside you, our trainers do exactly that in the classroom: timed drills, scored recordings and feedback on one weak spot at a time. Otherwise, choose your one weak spot today, record three answers and score them with the checklist. Then write the one line for your failure log before you close the laptop.
3 questions, about a minute
Check your understanding
- Mock interview
- Placement
- Data
- SQL
- Career
