Skip to content
Jobless Busters

Interview notebook

Jobless Busters

Practice for full-stack and AI engineering interviews: read model answers, test yourself, solve small coding challenges. Your progress stays in this browser.

Start with the first question Start practice

Questions

373

15 multiple choice

Answered

—

Answered correctly

—

Challenges passed

—

4 more for self-review

Continue learning

Good places to start:

Recent activity

Your answers and challenge runs will show up here.

Interview notebook

Full Stack / JS

From web basics to the full stack: HTML/CSS, JavaScript, React, Node.js, databases and HR.

All 176 questions

AI programme

AI & Agents Engineer

LLMs, RAG, agents, MCP and evaluation, with the backend and infrastructure around them.

All 197 questions

Learning paths

Ordered runs of questions, from a definition to a design decision. Stop at the first step you can't explain.

  • Networking and live data

    The path: “What is it?” → “What can the server send?” → “Does a chat need a WebSocket?” → “What happens after the connection drops?”. The point is being able to pick a simple solution, not claiming the newer transport always wins.

  • Queues, repeats and writes

    The path: “Why defer work?” → “Who carries it out?” → “What happens after a crash after a write?” → “How do you recognise the same operation?”. Only at the end do we connect two systems that one transaction doesn't cover.

  • Cloudflare, consistency and AWS

    The path: “What does the platform offer?” → “Why is an older copy sometimes enough?” → “When must you not rely on it?” → “How do you protect the last spot?”. The real variant of your campaign is asked about separately, after R01.

  • JavaScript and React

    The path: “What is the function waiting for?” → “Have the other operations started?” → “How do you handle them finishing in reverse order?” → “Why does the UI show a stale value?”. A good sequence for linking theory with short coding.

  • RAG and privacy

    The path: “How does the model know the document?” → “How do you find it?” → “Did the bad result arise before the model?” → “May this person be shown this passage?”. That matters more than choosing a fashionable vector database.

  • Agents and control over actions

    The path: “What does the model choose?” → “Who runs the function?” → “Do valid arguments grant permissions?” → “What if the connection fails after a write?”. The model proposes, but the system has to know the consequences.

  • Coding agents and quality

    The path: “What can the tool do?” → “What rights does it have?” → “What should it do now?” → “What evidence shows it works?”. When an agent writes a lot of code, you still need to be able to explain the most important part.

  • AI quality and your own projects

    The path: “What does a good result mean?” → “Is the system missing good things?” → “Who checked the judge?” → “What did you actually measure in your project?”. Don't replace a missing measurement with a nice-sounding number.