Mistake 1 — Optimizing problems solved per week
My first four months of prep were a treadmill: 25–30 problems a week, moving on the moment my solution passed. The result was that I could solve problems I had seen and froze on everything else. I changed to a review-first workflow: after each problem I wrote one sentence on the pattern, the trigger keywords, and the one trap I hit. I also re-solved each problem a week later without looking at my old code. That single change — forced recall — did more for my interview readiness than the previous four months of volume.
Mistake 2 — Grinding Hard problems too early
I spent my second month on LeetCode Hards because they felt impressive. This was backwards. Hards are mostly Medium patterns stacked together; without automatic Medium fluency, I could not even recognize the sub-problems. The fix was a 3:1 ratio — three Easy/Medium solves for every Hard attempt — and only touching a Hard once its underlying patterns were fluent. Interview performance is built on automatic pattern recognition, which only comes from volume at the right difficulty.
Mistake 3 — Never timing myself
I practiced untimed for months. My first mock interview fell apart because I spent 25 minutes on a problem I could have solved in 12 — no timer discipline, no sense of when to abandon a dead-end approach. From then on every practice session ran on a timer: 25 minutes for Medium, 40 for Hard, and I treated "time up" as a data point, not a failure. Learning to cut losses and switch approaches under the clock is a skill, and it only develops with a clock present.
Mistake 4 — Solving silently
I did almost all my practice in silence. When a real interviewer asks "what are you thinking?" mid-solution, silence reads as stuck. I forced myself to narrate every practice solve — stating my plan, the pattern I was considering, and why — even when solving alone felt unnatural. It took about ten sessions to stop feeling silly, and the payoff was immediate: interviewers stopped prompting me because the narration answered their questions before they asked.
Mistake 5 — Ignoring the final review week
Before my first loop, I used the last week to attempt new Hard problems, which mostly demoralized me. The week before an interview should be about consolidation: re-read your problem log, re-solve your marked weak spots, run two mocks, and stop. Fresh problems the night before an interview only add noise. The confidence that comes from reviewing a hundred problems you genuinely understand beats any last-minute discovery.