Prompt Patterns for Data Structures & Algorithm Problems

AI can solve almost any standard DSA problem in seconds — which is precisely the trap. If you let it jump straight to the optimal solution, you skip the reasoning that makes you a better engineer. The right prompt pattern uses AI as a tutor, not a cheat sheet.

1. Introduction

Whether you are preparing for interviews or trying to choose between a HashMap and a TreeMap for a real production decision, AI is at its most useful when it walks you through the reasoning. Asking for "the answer" gives you code; asking for "the approach, then the trade-offs, then the code" gives you understanding. This tutorial shows you how to structure both kinds of prompts.

2. The Concept Explained

Every algorithmic problem has four layers you can think through: the problem restatement, the approach choices (brute force, optimised, edge cases), the code, and the complexity analysis. A great prompt asks AI to walk through each layer in order — so you can stop and discuss at any point.

Think of the AI as a senior who has solved the problem a thousand times. The wrong question is "what's the answer?" — they would just tell you. The right question is "walk me through how you'd approach this if you'd never seen it." That's the prompt you want to write.

3. The Problem Without This Technique

The default prompt is to paste the LeetCode problem and ask for the solution. You get an optimised solution and no understanding.

solve this in python

Given an array of integers nums and an integer target,
return the indices of two numbers that add up to target.

4. The Solution

You are a calm DSA tutor. Walk me through this problem in four stops:

Problem:
Given an array of integers nums and an integer target, return the indices
of the two numbers that add up to target. You may assume each input has
exactly one solution and you may not use the same element twice.

Stops:

1. RESTATE
   - Restate the problem in your own words.
   - List inputs, outputs, and constraints I should clarify with an
     interviewer (sorted? duplicates? negative numbers?).

2. APPROACHES (no code yet)
   - List 2–3 approaches in order of increasing efficiency.
   - For each: rough idea, time complexity, space complexity, when it
     would be preferred over the others.

3. CODE
   - Implement the BEST approach in Python 3.11.
   - Use clear variable names, no one-letter names.
   - Add a 1-line comment above any non-obvious line.

4. COMPLEXITY & EDGE CASES
   - State final time and space complexity.
   - List the edge cases the implementation handles correctly.
   - List one realistic input where this approach would be a poor choice
     and explain why.

Pause between sections so I can ask questions. Don't move to the next
stop until I say "next".

5. Step-by-Step Breakdown

1. Frame the AI as a tutor, not an oracle. "Walk me through" produces teaching; "solve this" produces a solution.
2. Insist on restatement first. If the AI's restatement is wrong, the rest will be wrong. Catching it early saves time.
3. Ask for multiple approaches before any code. This forces explicit trade-off comparison — the heart of DSA reasoning.
4. Demand the language and naming convention. Single-letter variables are interview-acceptable but unreadable for learning.
5. Always ask for complexity and edge cases. A solution without complexity analysis is half a solution.
6. Use "next"-style pauses. Telling the AI to wait between sections turns it into a real conversation, not a wall of text.

Tip: Once you understand a problem, ask the AI to generate three variations of it (e.g. "what if the array is sorted? what if we want all pairs? what if we want pairs that sum to within K of target?"). Solving variations is how you build real algorithmic intuition.

6. Practice Exercises

List 3 access patterns where each option wins. Pick the right one for [your specific use case].

7. Key Takeaways