Yes, but for a different reason than before. AI can produce a sorting function; it cannot tell you whether your design will fall over at scale or whether the code it wrote is quietly quadratic. DSA teaches you to judge code, and many companies still test it. Learn the core patterns well, skip the endless grind.
What people tell me
I am a student and I have spent months on LeetCode-style problems. Now AI tools can solve most of them in seconds, and I keep reading that coding rounds will disappear. I am starting to feel that all this practice is wasted time and that I should only build projects with AI instead. At the same time, the companies I want to join still seem to have DSA rounds.
A composite of the messages behind this question, with personal details left out.
Key takeaways
- DSA is the vocabulary for reasoning about cost: time, memory and what happens as data grows.
- You need it to review and debug AI-written code, not only to write code yourself.
- Many companies still include problem-solving rounds, so it remains part of getting hired.
- Learn around fifteen core patterns deeply instead of chasing a problem count.
- Balance it with projects: roughly a third of your practice time is plenty for most people.
What DSA was ever for
Data structures and algorithms were never really about memorising how to reverse a linked list. They are about a way of thinking: what does this code cost, what happens when the input is a hundred times bigger, and which structure makes the common operation cheap. That thinking is exactly what you need when a tool hands you fifty lines of plausible-looking code.
AI changes who types the loop. It does not change the fact that someone has to know whether the loop is right.
Where DSA shows up in real work
| Situation | What DSA knowledge lets you do |
|---|---|
| Reviewing AI-generated code | Spot a nested loop over a large list, or a lookup that should be a hash map |
| A slow page or endpoint | Recognise that the problem is an algorithm, not the server |
| Designing a feature | Choose between a queue, a cache, an index or a sorted structure |
| Debugging | Reason about edge cases: empty input, duplicates, huge values |
| System design | Understand why databases use B-trees and why caches evict the way they do |
I see this often: developers who lean on AI without fundamentals ship code that works on test data and falls over in production. They cannot tell why, because they never learned to read cost.
It is still part of getting hired
Hiring processes change slowly. Many companies, especially larger product companies, still include some kind of problem-solving round, and some have moved to formats where you explain, extend or debug code rather than write it from scratch. Those newer formats test understanding even more directly. Treat DSA as one input to getting hired, not the purpose of learning it, but do not assume it has disappeared because a tool can solve textbook problems.
How to learn it now
The old approach was a problem count: 300, 500, 800 solved. That was never efficient and is even less so now. A better approach:
- Learn the core structures properly. Arrays, hash maps, stacks, queues, linked lists, trees, heaps and graphs. Know what each makes cheap and what it makes expensive.
- Learn patterns, not problems. Two pointers, sliding window, binary search, BFS and DFS, backtracking, dynamic programming, and a handful more. LeetCode Patterns by Sean Prashad and the NeetCode Roadmap group problems this way.
- Solve by hand first. Give yourself 25 to 40 minutes before any help. Struggling is where the learning happens.
- Then use AI as a tutor. Ask for a hint, not the answer. After solving, ask it to critique your complexity analysis or suggest edge cases you missed.
- Explain it out loud. If you cannot explain why your solution works, you have not learned it yet.
The DSA roadmap follows this order, and the DSA cheat sheet keeps the patterns and complexities in one place. VisuAlgo is useful when a structure will not click.
How much is enough
For most people aiming at a first or second developer job:
- Comfortable with easy problems and able to solve common medium problems in a reasonable time.
- Able to name the pattern within a few minutes of reading a problem.
- Able to state time and space complexity without guessing.
Beyond that, returns drop quickly unless you are aiming at companies known for very hard rounds, or you enjoy competitive programming for its own sake.
Keep it in proportion
The mistake is not learning DSA. The mistake is doing only DSA. A reasonable split for a student is about a third of practice time on problems and the rest on building, deploying and reading real code. Projects show you can ship; DSA shows you can reason. You need both, and AI makes the reasoning part more valuable, not less.
If you are still stuck
If you cannot solve problems at all yet, start with How do I start DSA when I cannot solve problems and never know which approach to use?, then read How many LeetCode problems are enough, and how do I practise them properly?. If you want a practice plan that fits your timeline, join Sefism and, as a member, book a 1:1 session.
Was this answer helpful?
Read next
- Fifteen Problem Shapes and the Move That Solves EachFifteen LeetCode patterns written as signal plus move: hash maps, two pointers, sliding window, prefix sums, heaps, backtracking, dynamic programming, BFS, DFS and Dijkstra, with a recognition table and a way to practise so they stick.
- Choosing a Data Structure Is Choosing What to Make CheapHow to choose the right data structure: twelve situations paired with the structure that makes the frequent operation cheap, a time complexity table, three questions to ask before storing anything, and the mistakes I see most.
- How to Use AI as a Companion, Not a CrutchHow to use AI for coding without losing your skills: the one rule that decides it, what the research says, where AI helps and where it quietly hurts, how students should learn with it, and how to verify its code.
Your situation is not quite this one?
Members get written answers to their own questions, a roadmap built for them and feedback on their projects. Early access is open: sign up and your account is active shortly after. Prefer to talk? A free call works too.