Skip to content
Tutorials Center
My learning

Topics

Lesson 1 of 77 Β· How Interviews Work

How Technical Interviews Work

Learn the stages of a software interview loop, what interviewers actually score, the main question types, and how this track helps you prepare for each.

A technical interview looks like a quiz, but it is really a short, structured conversation. Someone you have never met has about 45 minutes to decide whether they would like to work with you, and they must write down evidence for that decision. Once you know what they are writing down, preparing becomes much less mysterious.

This track is a question bank with explanations, plus runnable stand-ins that show how each idea works inside. This first chapter covers the process itself. Later chapters cover languages (JavaScript, Python, Java, C++), the web, databases, computer science fundamentals, system design, and the human side of hiring.

The stages of a typical loop

Companies differ, but most loops follow the same shape. Small companies may compress it into two conversations; large ones may run five or six rounds.

A typical interview loop

Recruiter call Motivation, salary range, logistics.
Technical screen One coding or concepts round, 30–60 minutes.
Onsite or virtual loop Several rounds back to back.
Decision and offer Interviewers compare notes.
1 / 2

The recruiter call is not a test, but it is evaluated. Be able to explain in one minute what you have built, what you want next, and why this company.

The technical screen filters candidates quickly. It is usually one coding problem or a rapid series of concept questions. This is where most candidates are rejected, so the basics (arrays, strings, hash maps, the core of your main language) matter more than exotic topics.

The loop repeats the same kinds of conversation with different people. Each interviewer writes feedback independently, so one weak round can often be recovered by strong ones.

What interviewers actually score

Interviewers rarely score "got the right answer". Most rubrics split the evaluation into a few dimensions, and a wrong final answer can still score well on most of them.

  • Problem solving. Did you break the problem down, try examples, and find a workable approach?
  • Technical knowledge. Do you understand your language, data structures and the tools you claim on your resume?
  • Code quality. Is the code readable, correct on edge cases, and reasonably tidy?
  • Communication. Could the interviewer follow your thinking? Did you ask good questions?
  • Collaboration. Did you take hints well and adjust, or defend a broken idea?

Notice that three of five are about how you work, not what you remember. That is good news: they are trainable.

When you are stuck, the interviewer is not only watching the clock. They are collecting evidence for "collaboration" and "communication". Calm narration of what you tried is real evidence.

Question types you will meet

Almost every question belongs to one of a handful of families. Naming the family tells you what the interviewer wants.

Type Example What they want to see
Concept "What is a closure?" A precise definition, an example, a pitfall
Predict the output "What does this snippet print?" You can trace execution without running it
Coding "Find two numbers that sum to a target" A correct, clear solution with complexity
Design "Design a URL shortener" Requirements, trade-offs, sensible components
Debugging "Why does this code leak memory?" A method for narrowing down the cause
Behavioral "Tell me about a conflict" A specific story with your own actions

How this track is organized

The chapters follow the order you are likely to need them: process first, then languages, then fundamentals, then design, then the human side, and finally projects and practice sets.

  • Each language lesson groups questions by theme (for example scope and closures) rather than listing them randomly, because understanding one idea answers ten questions.
  • The technology itself (a Java VM, a browser, a database server) usually cannot run inside this website. Real code for those is shown read-only, and every lesson also has runnable stand-ins, small programs that model the idea, and says honestly which is which.
  • The projects near the end build real tools such as a flashcard trainer and a mock interview runner, so you practice while you code.

A scoring rubric you can run

Here is a small stand-in for how a rubric might work. It is not any company's real system: it simply shows how dimensions become a decision, and why one weak area does not decide everything. Change the scores and see how the result moves.

JS A toy interview rubric

Press Run (or Ctrl + Enter) to see the console output here.

No output.

The "solid all round" candidate beats the one who solved the problem silently. Real committees also look for red flags, such as dishonesty or disrespect, which no weighted sum captures. The point of the model is the lesson: balance beats a single spike.

Preparing for the format, not just the content

Candidates often practice by reading answers. Interviews are spoken and timed, so practice that way: say the answer out loud, write code without autocomplete in a plain editor, and time yourself. The lessons on a framework for answering and thinking out loud train exactly that.

Check your understanding

Quick check Β· 4 questions

of 4 answered You scored / 4 Β· perfect!

  1. 1. Which stage is usually where most candidates are filtered out?

    The screen is a fast, cheap filter used on every applicant, so it rejects the most people. Basics matter most there.
  2. 2. An interviewer's rubric most often scores which of these?

    Rubrics split evaluation into several dimensions, so a candidate can score well even without a perfect final answer.
  3. 3. You cannot solve a problem in time. What is the best use of the remaining minutes?

    Clear narration gives the interviewer evidence of problem solving and communication, even without a finished solution.
  4. 4. Why does this track use runnable stand-ins?

    Most technologies need a server, VM or browser we do not have here, so small models show the idea while real code stays read-only.

Summary

  • A loop usually has a recruiter call, a screen, several onsite or virtual rounds, and a decision.
  • Interviewers score several dimensions: problem solving, knowledge, code quality, communication and collaboration.
  • Most questions fall into a few families: concept, predict-the-output, coding, design, debugging and behavioral.
  • Balanced performance beats a single spike, and communication is trainable.
  • Practice spoken and timed, not only by reading.

Next: A Framework for Answering Any Question