Andela found mismatched skills in almost 50,000 engineering postings

On 17 September 2026, HR Dive reported an Andela analysis of almost 50,000 postings for engineering roles. The analysis found mismatched skills and a proliferation of titles.

That is one firm's sample of engineering postings, as published that day. It is not a census of the whole market. It does not prove why any single employer wrote a req the way it did. It also does not, by itself, prove that AI caused the mismatch. The headline on the piece is broader than the finding quoted here.

The finding that is on the page is still worth keeping: in that engineering set, skills and titles were not the same object. Almost 50,000 postings is a large enough sample to notice.

A separate crawl from today#

Do not fold the Andela sample into today's PinchForth Job Index. They are different windows and different methods.

Today's Job Index crawl added 31 new roles. In that crawl, the public lead is cybersecurity (36 mentions). Also loud in the same crawl: OKRs (28), Ruby (18), SQL (17). Those are counts inside this crawl, not a share of all hiring.

Versus the 11 September snapshot (7-day window), javascript and kubernetes each show +11 mentions in the current top-skill list. Versus 19 August (30-day window), communication is +37 and ruby is +18. Those deltas are snapshot-to-snapshot inside the Job Index export. They are not the Andela study.

What to do with that#

If you are hiring or looking this week, keep two columns: the title line, and the skill list. The Andela sample says those two already came apart in a large engineering set. Today's crawl says cybersecurity still leads this board. Those can both be true without being the same story.

Rocket is the room where proof turns into intros. Bring the skill list, not just the label.