When we designed certification for Anested Skills, we started from a blunt question: what would make us trust a certificate enough to interview the person holding it? Recall of syntax trivia was not the answer. Nobody has ever shipped a product by remembering which array method mutates in place.
So our assessments are small, time-boxed builds. A candidate receives a specification — the kind a real client would actually write, ambiguities included — and ships a working project against it. Proctoring keeps the process honest; the spec keeps it realistic. The exam experience is deliberately closer to a working Tuesday than to a school test.
Grading is rubric-based and examines what production code review examines: does it work, is it readable, does it handle failure gracefully, and did the candidate make sensible trade-offs under time pressure? A working-but-messy submission and an elegant-but-broken one both learn something specific from their scores, because the rubric tells them exactly where they stand.
Certificates are verifiable by link, not by PDF. An employer clicks once and sees what was built, when, under what constraints, and to what standard. A certificate that cannot be verified in ten seconds is decoration, and the market treats it accordingly.
We put our own hiring where our rubric is: distinction-level holders enter Anested's own interview pipeline. That is the strongest signal we can send that we believe our own assessment — we are willing to staff our teams by it.
Students describe the exam as 'a hard but fair Tuesday at work'. That sentence is the entire design document. Skills certification should predict the job, and the only reliable way to predict the job is to sample it.
Need help with something we wrote about?
Our engineers answer support queries directly — hosting, databases, code, or anything from this post.