Hacker News new | ask | show | jobs
Show HN: findatechjob.dev – fed up with LinkedIn spam so built a job board (findatechjob.dev)
4 points by rornic 114 days ago
TL;DR: grew frustrated with boards filled with low quality ads and recruitment spam, so built my own instead.

Aggregates over 23,000 engineering roles from solid tech companies. Mostly startups and scale-ups all across the world.

Has a number of tech-specific filters with very fast faceted search interface. Plan to expand this.

I found my current role in the process of building this.

Let me know what you think.

2 comments

good work..

one suggestion.. all these page-content-rewrites might be fine for a magazine, but here the user is searching, probably comparing, probably hundreds of items. Mostly because of inadequate/impossible filtering, combined with quite tough market - the net has to be cast wider and wider. So, just give plain pages, with choosable page-size. SPA not exactly needed.

btw Remote-ness is highly volatile factor, cannot be booleanly relied.

example search: i am in Bulgaria, so anything there + anything ~~remote in EU + anything ~~remote in that GMT+2 timezone +/- 3h. Exclude anything from (list of shitty agencies). Now Read all those and check

- see if they are in my city or are enoughly remote (can be certain-country/ies-remote, or even certain city/ies-remote =~~hybrid), or pretending to be remote..

- see if they are written in or require a human language i do not speak

- and only then maybe check the actual role :/

in 3 months i have read like 5-25 of these per day, picked from ~10-100.. from 5 sites, eventually - if there were any.

i Agree that linkedin is full of shit, although they are the only ones that do have an exclusion operator "-something", and have somewhat working boolean filtering (X OR Y) AND (Z OR T).. eh, within their sea of shit :/

good to have some sane alternative, thanks

Appreciate the feedback and take all your points.

Location-based filtering and figuring out if a job is remote are tricky problems, as all the data sources I'm using represent it inconsistently.

Parsing location strings like "Multiple Location" and "100% Remote" or "Remote: United States | Europe | United Kingdom | Canada" is a bit of a ballache without latlon pair. The algorithm I've got right now needs some improvements.

sick what did you use to build it?
Rust and Svelte. Search powered by Meilisearch. All hosted across Cloudflare Workers and GCP Cloud Run, costs me virtually nothing to run.
Very cool!