> Rather than entirely shifting away from existing processes, teams should start to consider incorporating code reviews as a key step in the candidate selection process to identify those with a proficiency to effectively evaluating code.
They should indeed shift to code reviews to test the reasoning ability of the candidate's written code rather than already solved leetcode puzzles which is very easy for candidates to use LLMs to cheat. On last resort, I'd use leetcode but focused on the relevant language that I am looking for. (If I want to see the leetcode problem solved in Golang, I do not want to see you use Python to solve it.)
I would give the candidate a chance to skip the leetcode puzzles if they are able to demonstrate that they have contributed (at least two) to significant open source project(s) in the relevant technology in the job description and show an example code review of those commits.
This gives a public example of a candidate that is able to contribute to very high quality repositories that have been used by millions of developers and isn't afraid of public scrutiny or code reviews by the maintainers. Otherwise, I would give a standard leetcode task (hard) in the relevant language and code-review that in the second interview.
Eliminating 95% of AI frauds in one technical interview.
They should indeed shift to code reviews to test the reasoning ability of the candidate's written code rather than already solved leetcode puzzles which is very easy for candidates to use LLMs to cheat. On last resort, I'd use leetcode but focused on the relevant language that I am looking for. (If I want to see the leetcode problem solved in Golang, I do not want to see you use Python to solve it.)
I would give the candidate a chance to skip the leetcode puzzles if they are able to demonstrate that they have contributed (at least two) to significant open source project(s) in the relevant technology in the job description and show an example code review of those commits.
This gives a public example of a candidate that is able to contribute to very high quality repositories that have been used by millions of developers and isn't afraid of public scrutiny or code reviews by the maintainers. Otherwise, I would give a standard leetcode task (hard) in the relevant language and code-review that in the second interview.
Eliminating 95% of AI frauds in one technical interview.