Interviews aren't passed by the people who know the most. They're passed by people who can say what they know, briefly, clearly and under pressure. That's a skill, and it can be trained. This part covers tools for three kinds of prep: theory questions, coding tasks and system design, plus delivering answers out loud.
This is part 8, the last in the series. If you landed here first, start with the overview in part 1.
What gets tested
For a web developer, an interview usually has several parts:
- theory: JavaScript (closures, event loop, prototypes), TypeScript, the browser, HTTP, your framework;
- code: live coding or a take-home; for frontend often "build a component in 45 to 60 minutes", not just algorithms;
- system design: for middle and above; in frontend that means app architecture, state, caching, rendering;
- behavioural: how you work in a team, handle conflict, make decisions.
The higher the level, the more weight system design and behavioural questions get. More in the article on levels from junior to senior.
Theory and answering out loud
This is why ShortLesson exists. On ShortLesson you get:
- interview questions on JavaScript, React, SQL, Git and HTTP with worked answers;
- spoken answers graded by AI: you answer out loud for up to three minutes, the AI compares your answer with the points of a reference answer, shows what's missing and asks a follow-up. Pace and filler words are shown separately. The free plan includes 2 graded answers a day;
- spaced review: a question comes back after 7, 30, 90 or 365 days depending on how you answered (paid plans);
- cheat sheets for a quick refresh before the interview.
For general speaking practice there's Yoodli: roleplay scenarios with speech analysis. The free plan gives 5 scenarios; Pro is 8 dollars a month billed annually. It isn't about technical questions, but it shows clearly how you sound.
Coding practice
| Platform | Good for | Price |
|---|---|---|
| LeetCode | Algorithms and data structures, biggest library | Lots free, Premium is paid, check the current price on the vendor's site |
| NeetCode | Structured problem lists with video walkthroughs | Core is free, Pro is paid |
| GreatFrontEnd | Frontend-specific tasks: components, JavaScript utilities, frontend system design | Partly free, paid plan available, check the current price on the vendor's site |
| BFE.dev | "LeetCode for frontend", 600+ JavaScript, TypeScript and React problems | Mostly free, with premium content |
| Codewars, HackerRank | Warm-ups, short problems | Free |
A tip for frontend developers: don't bury yourself in LeetCode alone. In Ukrainian and many European companies, frontend live coding is more often "write debounce", "build an autocomplete", "find the bug in this component" than graph problems. GreatFrontEnd and BFE.dev are closer to that format.
Mock interviews with people
The closest thing to a real interview. And the most expensive.
- interviewing.io: anonymous mock interviews with engineers from big companies. Sessions are paid and not cheap; check the current price on the vendor's site. Worth it before an interview at a company you really want.
- Aced (Exponent's new name since 2026): system design and behavioural interview courses, coaching, an AI coach. The Aced family also includes Pramp (free peer mock interviews with other candidates) and Interview Cake.
- Hello Interview: strong system design material. They ended live human mock interviews on 31 May 2026; mocks are now AI-based, from about 10 dollars each. Premium is about 79 dollars a year.
- Big Interview: video course and behavioural interview practice, from 39 dollars a month.
A free alternative that works just as well: pair up with a developer friend and interview each other. Or ask a middle or senior you know to ask you 10 questions. Feedback from a real person in your market is often worth more than a paid platform tuned for FAANG.
Mock interviews with AI
Any AI assistant can play the interviewer. The prompt is in part 6. Pros: free, any time, as often as you like. Cons: AI can be too lenient and doesn't always catch inaccuracies. So compare against reference answers rather than trusting its score blindly.
If you have the connectors from part 7, the AI can build a prep plan straight from the posting in your mail and the invite in your calendar.
About AI hints during the interview
Some products listen to the interview in real time and show you ready-made answers. For example, Final Round AI and LockedIn AI advertise such "copilots" and modes that supposedly don't show up on screen share.
Why we don't recommend them:
- It shows. Eyes drifting aside, a pause before every answer, perfect phrasing that falls apart on the first follow-up. Experienced interviewers notice.
- Companies are responding. Some are bringing back in-person rounds and changing question formats because of these tools.
- You get a job you aren't ready for. Nobody gets through a probation period with hints.
- Reputation. If you're caught, people will hear about it. The industry is small.
The same technology is honestly useful before and after an interview: practice, answer reviews, a review plan.
A one-week plan before an interview
- Day 1. Break down the posting (prompt in part 6) and pick 5 to 7 topics.
- Days 2 to 4. For each topic: a cheat sheet, 10 interview questions, 3 to 5 answers out loud.
- Day 5. Two timed live coding tasks on GreatFrontEnd or BFE.dev.
- Day 6. A mock interview with a friend or AI, plus 3 stories for behavioural questions.
- Day 7. Light review of weak questions, rest.
A vacancy mode that will build a plan like this automatically from the posting is in development. More in part 4.
End of the series
That's the whole stack: job sources, a tracker, CV and ATS, LinkedIn, AI assistants, MCP connectors and interview prep. You don't need it all at once. Take the minimal stack from part 1 and add tools when a specific need shows up.
