Jev it up guys!

Search was my weakest spot. It was plain MySQL fulltext, so short or tricky terms were broken. "Go developer", "AI engineer" and "C# developer" returned 0 or garbage results. Adding filters like seniority, salary and location made it worse, you often got an empty page.
This week I shipped AI search with Jev:
Just type "senior react dev in Berlin $150k with equity" or "junior python engineer London 100k" and it works.
How it works under the hood:
Fast regex extracts tech tags, city, seniority and salary. I built a 120 term taxonomy so React.js, Golang, C#, Go and AI all map correctly.
Jev System One (hosted on Cloudflare Workers AI) reads the fuzzy intent: benefits, regions, pay interest, recency.
If strict filters give less than 5 hits, it auto broadens in tiers: city to region to worldwide, then relaxes seniority and salary. You never hit a dead end.
Impact so far: Senior Angular in Berlin went from 2 to 165 results. AI engineer went from 0 to 3,050. Go and C# finally work.
It is live on all 3 boards now for Pro users, and free search also got tag matching so it handles Go and C# better.
Curious: if you run a directory or marketplace, would you trust auto broadening, or would you rather show "0 results" to stay strictly accurate?