Touheed Khan

Posted on Jun 28

We Pulled Together 4,628 Technical Interview Questions Across 8 Stacks. Here's What the Collection Looks Like.

#career #interview #webdev #programming

We built our prep library by working through established syllabi, official docs, and the topics that keep coming up in technical interviews. Then we counted the whole thing. Here's the honest breakdown — and how we'd use it.

Building an interview-prep library is a strange exercise: you spend months deciding what's worth knowing, and you never step back to look at the shape of what you've made. So we did. We counted every question we've curated, where it sits, and how hard we rated it.

A note up front, because it matters: this is a breakdown of our collection, not a survey of real interviews. We didn't scrape thousands of interview transcripts. We assembled these questions the way a good study guide gets made — from public syllabi (our SQL set, for example, follows the roadmap.sh outline), official language and framework docs, and the topics that recur across technical interviews. So the numbers below describe what we chose to cover and how we tagged it — useful as a study map, not as a statistical claim about the industry.

With that said, here's what 4,628 questions look like.

The shape of the collection

4,628 questions, each with a full written answer and code
265 topic pages (e.g. "Python decorators," "SQL window functions")
8 stacks: Python, Java, JavaScript, React, SQL, Spring Boot, FastAPI, .NET

How it breaks down by difficulty

Every question carries a difficulty tag — our own editorial call, applied consistently across the library:

Difficulty	Share
Medium	~55%
Hard	~23%
Easy	~22%

The middle is where most of the library lives, and that matches how we think about prep. Easy questions are recall — you know that is compares identity and == compares value, or you don't. Hard questions are stretch — nice to nail, rarely the thing that sinks you. The medium tier is the working part of the exam: "walk me through how the event loop schedules this," "why does this GROUP BY return the wrong count." You can reason these out, but only if you understand the mechanism. That's our opinion on where to spend study time — offered as a study heuristic, not a law.

Where the questions cluster, by stack

Stack	Questions
Python	797
Java	778
JavaScript	744
React	670
.NET	450
Spring Boot	428
SQL	417
FastAPI	344

Two honest forces drive these numbers, and it's worth separating them:

Genuine surface area. Python, Java, and JavaScript are decades-old languages with deep, well-documented sets of "gotchas," so there's simply more worth asking about.
How far along our coverage is. Python and Java are our most built-out stacks; FastAPI is one of our newest. So a smaller count partly reflects a younger, narrower syllabus — not that the stack is "easy."

We're calling that out because the alternative — inventing a tidy story about why SQL "should" have fewer questions — would be dishonest. The counts will shift as we keep building, and we'll update this post when they do.

How we'd actually use this map

Spend most of your time in the medium tier. It's the bulk of the library and, in our experience, the bulk of the interview.
Go deep on one stack, not wide on five. 797 Python questions isn't a weekend. Depth in your primary language beats thin coverage everywhere.
Treat narrow stacks as high-stakes. When a topic like SQL has a small surface area, every question carries more weight — there's nowhere to hide on a join.

It's not just a list of questions

A pile of 4,628 questions is a reference, not a study plan. So we wrapped the collection in three things that turn it into one:

Roadmaps — every stack has an ordered path from fundamentals to advanced, so you always know what to study next instead of staring at a wall of topics.
Flashcards — mark any answer "known" and the concept drops into a per-stack deck you can drill for fast recall.
Quizzes — test yourself on a stack before the interviewer does, and see exactly where the gaps are.

Progress is tracked per device as you mark answers known, so the roadmap fills in as you go and the "continue" button always points at your next topic. Same 4,628 questions — three different ways to work through them.

Browse the actual questions

Everything's organized by stack, each linking straight to the answer with code:

Follow a roadmap instead

Prefer a guided path over a question list? Each roadmap orders the same questions from fundamentals to advanced:

We'll re-run this as the library grows. Curious whether the ~55% medium split holds once we're past 10,000 questions.

DEV Community