A Better Way to Technical Interview

We’re tired of technical interviews that rely on puzzles for reliability. We want creative, bold thinkers.

We’re previous FAANG employees and startup founders and have interviewed hundreds of candidates. We’ve found some of our best candidates couldn’t do leetcode — and never would’ve passed at Big Tech — but were some of the most consistent problem solvers.

We learned collaborative, open-book interviews refuse to let strong founders fall through the cracks. Here’s how to replicate & implement it.

Use open-ended problems with multiple solutions

• The candidate is presented with a problem that can be solved in many different ways but also given exact requirements to guide to a solution
• Having a base problem that gradually gets built upon is a good type of problem to present
• The proctor has questions to ask during pairing and tips on where to align with the candidate and steer them toward alternative implementations
• A general rubric ranging from performance to readability to eliminate subjectivity

Pair programming

• The candidate should be informed of the type of problem they will be solving and the language they will be using (if you require a preference)
• The candidate should share their screen with the interviewer, this way you can assess their daily work environment, general efficiency, and usage of tools
• Your proctor should work on the problem with the candidate, not just answer questions
• The proctor and candidate should have a back-and-forth discussion of the problem, this isn’t a one-way street

Usage of all resources

• Allowing the candidate to use all resources available to them including AI-assisted tools
• Allowing the use of all libraries but with the caveat they must explain their reasoning for why they chose that specific library

As long as your prompt is truly open-ended (no “find all the substrings in a string!”), AI won’t be able to derive complete answers. These types of prompts will inhibit the problem from being solved in a single pass without human instruction. Oftentimes, there’s a library that fits the niche of a specific problem domain, and this should be no different in an interview.

Problems should be robust enough to encourage the candidate to throw everything they have at it.

Your next hire will be a core part of the team, and you need to know how they approach problems and tackle solutions. Treat your interviews as if this candidate was already on your team.

About Interviews

Interviews is a library of real-world programming scenarios to use as technical interviews. We offer both free and paid technical interview scenarios, learn more and how to join here.

Got some questions for us? contact@interviews.so