I recently went through an interview process for a full-stack position, and I'd like to share my experience. The process included a screening with a recruiter, where I was asked some basic JavaScript questions, and a live coding session with the development team.
For the live coding session, I received an archive containing a React app repository and was asked to add logic and components using hooks (controlled inputs, state management, context, etc.). I managed to complete about half of the requested tasks within the given timeframe. Here are the reasons why:
1. I was nervous due to my limited live coding experience, which significantly affected my speed of thinking.
2. The IDE autocomplete and linting were disabled. This meant that any import/export typo or a linting error had to be figured out from the compilation errors in the terminal or browser console. The same applied to API methods and standard loop blocks, which I had to recall from memory. This significantly slowed me down. I even voiced my frustration about the lack of autocomplete a few times and received feedback that I was "too dependent on autocomplete" LOL
Ultimately, I didn't get through because I didn't complete all the tasks on the list. However, I don't regret it, as I find this interview approach somewhat unfair. Most likely, the top candidates would either be those who game the system by applying together (where one collects the task and the other learns the solution beforehand) or simply ask recent candidates on Blind for the tasks, allowing them to complete everything in 15 minutes.
**Pros:**
- The interviewers were friendly. One interviewer mostly stayed quiet in the background. The Director of Engineering was very nice, often pointing out typos (like a missing bracket after API method calls) and highlighting existing template code for the functions I was about to implement.