I used to think that LeetCode problems were just a way for interviews to filter candidates from a huge candidate pool.

After all, once you get selected, you will probably be building a website, or running an ML model, or building an app or something, right?

So surely, building full-stack products is greater than DSA, right?

At least that's what I thought. But once I met a problem myself, I realized the importance of DSA.

I was trying to build a voice agent solution for a hospital. I made a prototype, and it worked.

So yeah, I successfully built a product.

So the cost of running the voice agent was like 2 rupees per minute, low, right?

But assume you have to run it for like 30k minutes per month, so the cost is now 60k rupees.

But had I made a slightly better data structure, I could just make the cost be like 1 rupee per minute or something (number exaggerated).

1 rupee might seem like a very low number, but when we scale (which all production-grade software will), that cost becomes very large soon.

So if you're studying, do not choose between DSA or building full-stack products. You actually need to do both.