Nearly three years ago, I wrote about the difficulties of finding programming roles without involving recruiters. The problem primarily is that high-quality direct roles are mostly pushed aside in search engine results in favour of highly cross-linked agency noise.

A simple solution to this is a crawler system that specifically indexes no-agency roles from their original locations and displays them, on a single website, with the usual search facilities. It would be rather good if jobs are snap-shotted, so we can see whether an advert has changed. A non-intrusive advert on the page would provide the funding for the site. Location data could be obtained from the reverse geocoding of postcode(s) on a contact page.

This system would need to parse (X)HTML and, per site, understand where the job title and other meta items are stored. Some jobs may be presented in iframes, which the system would also need to parse correctly.

Update 23rd June

This is proving to be an interesting problem, so I’m putting down a bit of code for this.

Since this is a sort of aggregator, some might argue it’s in the same niche as sites like indeed.co.uk and careerjet.co.uk – both source a variety of job feeds and show them on a searchable site. However, obtaining roles from third party lists causes duplication (not fully weeded out on Careerjet at least) and – major bugbear – each site is around 80% recruiters. Neither records much metadata about each role – accurate location data especially would be a huge win. In short, current market offerings can be improved on considerably.