We've been rethinking how companies assess software engineers. Here's what we're building.

One thing I've been thinking about while building Mindfries is how difficult it is to evaluate actual engineering ability.

We often assess candidates through coding questions, but real engineering involves much more: understanding unfamiliar codebases, debugging, building features, designing systems, managing infrastructure, and making technical decisions.

So we're trying a different approach with Mindfries.

We're building an AI-powered engineering assessment platform that focuses on practical technical work across different disciplines:

• Software Engineering (Backend, Frontend, Full-Stack)

• Cloud & Infrastructure (Cloud, DevOps, Platform, SRE)

• AI & Machine Learning

• Data Engineering

• Security, QA and specialized engineering

Candidates can work with codebases, use a development environment, execute commands, test solutions, and demonstrate their engineering approach.

We're also exploring system design assessments, AI-assisted evaluation, rubric-based grading, and conversational interviewing through Tanu, our AI interviewer.

We're still early, and there's plenty we're learning along the way.

I'd genuinely like to hear from other makers and engineering leaders here:

What's one engineering skill you think traditional technical interviews consistently fail to measure?

And if you've experimented with practical assessments in your hiring process, what worked and what didn't?

We're building Mindfries here:
https://mindfries.com