InterviewKit

LeetCode is dead, let Codex conduct technical interviews

2 followers

LeetCode is dead, let Codex conduct technical interviews

2 followers

I built InterviewKit because technical interviews should reflect how engineers actually work with AI. Candidates use Codex to complete a real task, but they still have to drive the process: choosing an approach, making trade-offs, directing the code, testing it, and dealing with failures. Employers see more than the final output. They get a clearer picture of the candidate’s judgment and how they work.

Launch tags:Career•OpenAI Day

Launch Team

Migma AIEmail marketing from Grok Bot

Promoted

OpenAI

OpenAI Day

How did GPT-5.6 change the ambition or scope of what you shipped?

Bone Health Tracker

Maker

GPT-5.6 helped turn InterviewKit from a small experiment into a complete product I could launch. Instead of stopping at the assessment flow, I shipped the employer dashboard, candidate plugin, privacy-preserving evidence capture, isolated evaluation, deployment setup, documentation, and launch assets. The biggest change wasn’t just speed. It helped me work across product design, engineering, verification, and presentation while keeping the whole experience coherent.

Report

2mo ago

Bone Health Tracker

Maker

📌

Hi Product Hunt 👋 I built InterviewKit because technical interviews were starting to feel disconnected from how software is actually being built. AI tools are becoming part of everyday engineering, but many interviews still ask candidates to pretend they don’t exist. I wanted to explore a different question: Can this person use AI with good engineering judgment? During an InterviewKit assessment, the candidate works with Codex on a real engineering task. They decide the approach, make the trade-offs, direct the implementation, choose what to test, and respond when something fails. Codex follows their direction, but it doesn’t coach them, suggest the right answer, or quietly rescue weak decisions. The candidate still has to drive the work. InterviewKit captures privacy-preserving evidence of that process so employers can evaluate more than the final code. Candidate work stays on their device until they explicitly choose to submit it, and submitted projects are evaluated in an isolated environment. This is an early pilot, and I’m still learning what an effective AI-native interview should look like. I’d especially love feedback from founders, engineering leaders, recruiters, and developers: What would you want to understand about a candidate’s ability to work with AI?

Report

2mo ago

Reviews

OpenAI

OpenAI Day

How did GPT-5.6 change the ambition or scope of what you shipped?

Bone Health Tracker

Maker

Report

2mo ago

Bone Health Tracker

Maker

📌

Report

2mo ago