The week before an interview is for making sure what you already know is instantly retrievable under pressure, not for learning anything new. The mistake almost everyone makes is treating the final week like prep week two: grinding fresh problems, panicking at gaps, and walking into the interview tired instead of sharp.
You cannot re-solve 370 problems in seven days, and you shouldn't try. Revision is a different skill from learning, and it has its own method: spaced repetition, fast recall passes, and timed re-solves of only the patterns you're shaky on.
The core idea: recall beats re-solving
Re-solving a problem you already cracked is mostly wasted effort; you're rehearsing a win. Revision should spend its time where retrieval is weakest. So most of your final week is a fast pass, not coding, built on three moves:
- Read the statement. Ten seconds, no more.
- Guess the approach out loud. Name the pattern, the data structure, the complexity. "Sorted array, pair sum: two pointers, O(n)."
- Verify against your one-line note. If your guess matches, mark it solid and move on in under a minute. If it doesn't, that problem goes on the re-solve list.
This read-guess-verify loop lets you cover 40-50 problems an hour. You proved you can code weeks ago. This pass checks that you can still recognise the shape fast, which is the part that decays.
Revise your weak spots, not the whole sheet
The honest constraint: you have time to deeply re-solve maybe 30-40 problems this week, not 370. So the whole game is choosing the right 30. Those are the problems you flagged, peeked at the solution for, or got stuck on the first time. If you tracked your prep honestly, this list already exists. DSA Tracker tracks your progress and runs AI skill analysis as you go, so your weak spots are already on record. Your revision list builds itself instead of relying on memory of what felt hard a month ago.
If you didn't track anything, spend the first hour doing the read-guess-verify pass across your sheet and writing down every problem where the guess failed. That failure list is your week.
The day-by-day plan
| Day | Focus | What you actually do |
|---|---|---|
| 1 | Triage | Fast pass across Tier-1 topics; build the weak-pattern list. |
| 2 | Weak pattern A | Timed re-solve of 6-8 flagged problems in your weakest pattern. |
| 3 | Weak pattern B | Same, for your second-weakest pattern (usually DP or graphs). |
| 4 | Company top 50 | Filter to your target company's recent questions; read-guess-verify all of them. |
| 5 | Re-solve misses | Timed re-solve of anything you missed on Day 4. This is the highest-yield day. |
| 6 | Mixed mock | Two timed problems blind, narrating out loud. Simulate the real thing. |
| 7 | Light + rest | One fast pass over your one-line notes. Stop by evening. Sleep wins interviews. |
Timed re-solves: 25 minutes, hard stop
When you do re-solve, do it under a clock. Give yourself 25 minutes per medium problem and stop dead when time's up, even mid-solution. The interview is timed; your practice should be too. Untimed re-solving builds false confidence: you feel fluent because you had unlimited time, then freeze when the clock is real. After the 25 minutes, read the optimal approach, update your one-line note, and move on.
The company top-50 filter
Generic revision is fine, but the final 48 hours should narrow hard. Pull your target company's frequently-asked list (most are public on the usual platforms) and run the read-guess-verify pass over those specifically. Companies reuse questions and lean on particular patterns; a fintech backend loop looks nothing like a product-company frontend loop. If you know the topics that show up most and overlay the company's tendencies on top, you walk in having rehearsed the most likely material instead of a random sample.
What not to do
- Don't start a new sheet. Switching sources in the final week resets your sense of progress and teaches you nothing.
- Don't chase hard problems you've never seen. A new hard problem you fail at on Day 6 only dents your confidence. Stick to problems you already know.
- Don't skip sleep to cram. A rested brain recognises patterns faster than a fried one. The last night is for rest, not Tier-3 trivia.
If you're reading this and the interview is further out, you're better off building the foundation first; see the 90-day placement plan and come back to this revision system in the final week. Last-mile revision only works when there's something to revise. Track your weak spots as you go, and the final week stops being a panic and becomes a checklist.
Related on DSA Tracker
Originally published on DSA Tracker, a free site to learn, practise and revise DSA with an in-browser compiler, aptitude practice and mock interviews.
Top comments (0)