Interviewing
Mock interviews
Mocks and recordings fix communication failures that silent grinding never surfaces. Grade process, not vibes.
Mocks convert knowledge into interview performance. Past ~100 well-reviewed mediums, another 50 problems helps less than three hard mock interviews.
Watching a recording of yourself is uncomfortable and disproportionately useful: you will hear rambling, silent coding and missing test cases.
Walking through a mock block
Do at least:
- 3 coding mocks (peer or paid)
- 2 system design mocks
- 1–2 behavioral mocks
Record yourself once.
Peers are free and uneven. Give them a rubric: did you clarify, did you test, did you analyze complexity, did you drive design? Swap roles so you learn to evaluate.
Paid platforms (Interviewing.io, Pramp-style peers, Hello Interview coaches, Exponent) buy consistency and feedback quality. One good coach session that names your failure mode is worth five polite peer sessions.
Self-mocks: open Excalidraw or a blank editor, set a 45-minute timer, speak aloud as if someone is there. Then compare to a known solution.
Grade yourself on process, not vibes:
| Signal | Pass | Fail |
|---|---|---|
| Clarifying | Constraints and examples before code | Coding in 90 seconds |
| Approach | States brute force then improves | Jumps to clever idea that breaks |
| Communication | Narrates intent while typing | Silent for long stretches |
| Testing | Walks examples and edge cases | "Looks good" with no dry run |
| Recovery | Uses hints and resets cleanly | Panics or argues with interviewer |
Keep a spreadsheet: date, type, time, pass/fail/partial, one failure mode. Review weekly. If "silent coding" appears three times, the next block is narration drills, not more DP videos. If design mocks die in requirements, drill clarifying questions for a week.
Day-of
Sleep and logistics beat last-minute grinding. The night before, skim your story bank titles and one design skeleton — do not open ten new LeetCode hards. For virtual onsites: wired connection if possible, charged headset, empty desktop, water, and a backup phone hotspot.
During the day, treat each round as independent. A rough morning coding round does not doom the afternoon design round unless you narrate your doom into the next interviewer's first five minutes. Reset between sessions.
Readiness is a distribution. You will still miss some loops after a strong prep block — question variance and interviewer calibration are real. The plan's job is to raise your pass probability, not to guarantee a single company's yes.
Common mistakes
- Polite peer mocks with no rubric. You leave feeling fine and learn nothing.
- Skipping the recording. The failure mode is almost always audible.
- Grinding new problems instead of drilling the failure the spreadsheet already named.
Check yourself
Your last three mocks show silent coding. What is tomorrow's practice?
Narration drills — talk while you type — not another DP video.
You have time for either 40 new mediums or three hard mocks. Which?
The mocks. Past a solid base, mocks beat volume.
Key takeaways
- Mocks and recordings fix communication failures silent grinding never surfaces.
- Use a rubric. Grade clarifying, approach, narration, tests, recovery.
- Drill the failure mode the spreadsheet named.
- Day-of: sleep, logistics, reset between rounds.
Further learning
Next, open the pipeline only when mocks say you are ready: batching applications.
A full software-engineering mock interview so you can hear strong communication:
Related
Interview preparation strategy
How loops work, what each round measures, and a 6–10 week prep plan that keeps coding, system design and behavioral on track.
Interview prep timeline
A 6–10 week plan with overlapping tracks — coding first, design overlapping, behavioral from day one, mocks last.
What interview rounds measure
Coding scores process under time. Design scores decomposition and trade-offs. Behavioral scores ownership. The debrief scores your level from all three.