jobs.devinl.im

01 Apr 2025

Overview ⤴

So this is my first blog post ever (hooray!), and my aim here is to concisely write about the motivation, general architecture, and design decisions behind this site.

First, the idea:

1. What is it?
   • It’s a site that scrapes jobs directly from company portals.
2. Why?
   • I’m about to graduate with my Masters in Data Science soon, and it’s generally a good idea to apply to recently posted jobs over old postings. However, your options to do this are usually quite cumbersome:
     • You can manually checking each company’s site everyday.
     • Some sites let you setup notifications individually.
     • You can download browser extensions that checks for page change on a timer.
     • There was actually one site I found that seems to be doing this: JobRadar, but they’re missing a lot of sources.
3. How?
   • Historically, the hard part in building something like this are usually the scrapers. You’d handcraft a scraper for each company’s site, which will eventually break and need updates. As you scale the number of sources, it becomes a neverending cycle of creating and updating parsers.
   • Instead of doing that, we can now just use LLMs to identify and extract relevant information to our desired schema -> no more building parsers!

…more writing to come :)