Job family
Assessing software engineering
Roles that build, test and maintain software, judged on code they can defend under questioning.
Software engineering is the rare high-volume family where the job and the assessment can be the same activity. The work product is text, produced alone at a keyboard, against a specification someone else wrote and did not finish writing. Hiring volume is substantial and, unlike most of the largest occupations, genuinely growth-driven: BLS projects 174,700 net new software developer jobs in the US between 2025 and 2035, within a group of 1.9 million, with around 106,100 openings a year. Eurostat's vacancy data shows the same occupation running roughly three times the all-occupation EU vacancy rate even after a post-2021 cooling of 1.5 percentage points.
The thing that has changed about screening this family is that producing correct code is no longer scarce. A candidate with an assistant open in another window can hand you a working solution to almost any self-contained puzzle. What is still scarce, and still predictive, is whether they understand the code they submitted: why that data structure, what happens when the input is empty, which part they would rewrite first and why, what the failure mode is under concurrency. The gap between a good and a bad engineering hire shows up in the maintenance tail — the edge case they did not consider, the test they did not write, the migration they wrote as a one-shot script — and none of it is visible from a green test suite alone.
That is why the design for this family pairs a monitored sandbox with a spoken follow-up about the candidate's own submission. The sandbox captures how the work was actually done: the order of edits, where time went, what was tried and abandoned. The interview then asks the candidate to explain and extend their own code, which is a question that only has a fluent answer if the reasoning was theirs. The correct posture is not to ban tools — it is to accept that tools were used and to assess the judgment that survives them. That is also the argument against the status quo of a timed algorithm puzzle: it screens for recall of a competitive-programming idiom that most of the 1.9 million people in this occupation will never use, while leaving code review, debugging someone else's mess, and honest estimation entirely unmeasured.
Why this work can be assessed
The deliverable is a text artefact produced at a keyboard, so a monitored sandbox reproduces the job almost exactly — and a follow-up interview about the candidate's own code separates authorship from assembly.
These roles hire on different signals
Each of these is assessed differently. An assessment that treats them as one job measures whichever of them it was designed for.
Roles in this family
Sources
Every figure on this page is traceable. Where a claim could not be sourced it is stated qualitatively instead.
- US Bureau of Labor Statistics, Occupational Outlook Handbook, Software Developers, Quality Assurance Analysts, and Testers, 2025, https://www.bls.gov/ooh/computer-and-information-technology/software-developers.htm
- US Bureau of Labor Statistics, Occupational Outlook Handbook, Occupations with the Most New Jobs 2025-2035, https://www.bls.gov/ooh/most-new-jobs.htm
- Eurostat, Job vacancy rate: main ups and downs by occupation, 11 February 2026, https://ec.europa.eu/eurostat/en/web/products-eurostat-news/w/ddn-20260211-1
Hiring for one of these? We build the assessment for the specific role, run it under your brand, and return a ranked list with the evidence behind every score.
Book a walkthrough