Our Take-Home Test Allows AI. Here’s Why.

20 Oct 2026 · 0 min

Watch on YouTube

0:29

In this brief presentation, Angelina Yang explains why her organization permits the use of AI in take‑home coding assessments. She advises candidates to choose a text‑focused problem area, build a simple demo that goes beyond merely replicating existing solutions, and incorporate a meaningful improvement. The core evaluation shift is from asking candidates to write code from scratch to assessing their ability to direct, validate, and critique AI‑generated code, ensuring they can oversee and trust the output rather than merely produce it themselves.

We try and give them a take-home and we just allow everything. We don't tell them, don't use AI. Pick a part of textual that interests you and then try to make your own demo, like simple version. Don't just recreate it, but also make some cool improvement that makes everything much more interesting than that improvement that they made over a solution that we assumed was just like 100% cloud-coded. Ultimately, we feel like with coding, the intelligence has stopped being, can you write the code, to can you direct it and can you check its work and are you doing the right thing?