Avri Simon

Founder/CEO @ Eval-X. 1,000+ technical interviews as a 5x CTO. Building how you evaluate engineers in the AI era - we measure AI collaboration, not cheating.

[Avri Simon @AvriSimon]

I ran 1,000+ technical interviews as a CTO.

Then AI broke every one of them. LeetCode, take-homes, trivia - all solved by ChatGPT in seconds.

So I stopped testing what AI can do, and started measuring how engineers actually think and work with it.

That's why I built Eval-X.

[Avri Simon @AvriSimon]

@KomalWCode Does your current loop test any of those?

[Avri Simon @AvriSimon]

@SanDiegoKnight That means moving beyond resumes and LeetCode, and actually evaluating how engineers think and build.

[Avri Simon @AvriSimon]

@avrldotdev That gap is enormous. And you can surface it in 45 minutes if you set up the eval correctly.

[Avri Simon @AvriSimon]

@code_bytein That's the signal: can you explain the 'why' behind the 'what'?

[Avri Simon @AvriSimon]

@AIGuideHQ The question I now care about: what does a candidate say when the AI output is subtly wrong?

[Avri Simon @AvriSimon]

@emollick That's a harder eval to design. Most companies haven't updated their interview loop yet.

[Avri Simon @AvriSimon]

@realpython Being able to explain the tradeoffs and potential failure points of a solution is a much stronger signal than just getting to the right answer quickly.

[Avri Simon @AvriSimon]

@vibeonX69 The tool doesn't fix the judgment problem. It just makes it more visible.

[Avri Simon @AvriSimon]

@smitadeshmukh That means moving beyond resumes and LeetCode, and actually evaluating how engineers think and build.

[Avri Simon @AvriSimon]

@its_shreeja21 Does your current loop test any of those?

[Avri Simon @AvriSimon]

@RokoMijic That's a harder eval to design. Most companies haven't updated their interview loop yet.

[Avri Simon @AvriSimon]

@srishtiberi The question I now care about: what does a candidate say when the AI output is subtly wrong?

[Avri Simon @AvriSimon]

@kmeanskaran That's the signal: can you explain the 'why' behind the 'what'?

[Avri Simon @AvriSimon]

@ajay_2512x That gap is enormous. And you can surface it in 45 minutes if you set up the eval correctly.

[Avri Simon @AvriSimon]

@yongfook That means moving beyond resumes and LeetCode, and actually evaluating how engineers think and build.

[Avri Simon @AvriSimon]

@jhylee95 That's a harder eval to design. Most companies haven't updated their interview loop yet.

[Avri Simon @AvriSimon]

@campbellclaret @CountBinface @RestIsPolitics You don't know what you can't measure.

[Avri Simon @AvriSimon]

@vibhor376 Does your current loop test any of those?

[Avri Simon @AvriSimon]

@SCR01111 The question I now care about: what does a candidate say when the AI output is subtly wrong?

[Avri Simon @AvriSimon]

@GithubProjects That's the signal: can you explain the 'why' behind the 'what'?