Are LeetCode-style interviews are more fair and objective?

There has been a growing trend of companies using take-home assignments to assess candidates for screening. A few of them takes it to the next level by converting their onsite rounds into a 5 hrs long coding project, e.g. use this mock to build a mini web app.

Initially, I was quite happy to welcome this trend. No LeetCode-ing, use my favorite tech stack, sharpen some fundamentals, and learn a new Web API or two. It didn't end there, I've received 4 offers from companies that used a combination of take-home exercise and code-a-project during onsite.

Fast forward to current date, I'd discourage this approach of interviewing a candidate. Firstly, the amount of time involved per exercise is impractical, think 3+ hrs per take-home exercise. And more importantly, any two teams could have entirely different opinion about the same piece of code/architecture. In a limited time setting, candidate can focus on a subset of key aspects like functional spec, clean code, tests, performance, or pixel perfection. But it's difficult to gauge what is important to the team.

I wouldn't want to criticize something without giving a few alternatives.

Option 1. LeetCode-style interview is difficult to prepare, but it is more objective and fair when compared with code-a-project styled interview.

Option 2. For frontend engineering roles, there are well defined focus areas. Onsite interviews should assess each of these areas separately rather than clubbing them into a single mini-project. Example:

Round 1. HTML/CSS, tests solid foundational knowledge. Round 2. JavaScript, tests solid foundational knowledge. Round 3. Framework, pair program a widget. Round 4. Special Topics: accessibility, networking, performance, security, testing etc, Pick 1-2 topics and go deep. Round 5. System Design for experienced candidates/CS Fundamentals for lesser experienced candidates.

#frontend #interviews