The idea of this demonstration is to show that it's possible to meet modern user experience expectations of typing / saying what they're looking for, doing away with the need to fill in forms - regardless of how short they might be.

This demonstration is all about code. The data was taken from Kaggle, illiminating the possibility of designing an efficient data structure, although it obviously provided the advantage of having a reasonably large, real world data set.

Near-natural Language

The code was written to allow the input of near-natural language search terms in a single post and prioritises short search times for lower yielding results. What this delivers is longer search times for single word queries, but very quick performance for more naturally structured requests, for example: "I'm looking for an engineering job in the south east"

Of the 50,000 UK-based jobs in the spreadsheet, slightly more than 20,000 had specified job requirements and it was those that were chosen for inclusion.

So what you're getting with the search results / performance is an algorithm going to work on a spreadsheet of data that's been imported to a database.

These steps could be taken to improve performance

If you're interested in working together on your project, I'd be delighted to hear from you.