Jabbs

I run a job search site (UnlistedJobs) which monitors about 200k company website career pages. So my data comes from that and I've normalized it to look at the same group of companies (for analysis sake). Ah yea I've seen that Indeed chart circulating.

I'm in the process of putting together a market analysis tool with my data since it isn't biased to one platform (not saying that Indeed data is wrong though). Mine will only go back to the beginning of 2025 tho

Yea totally agree, would like a 10+ year chart. I only had good data from the beginning of 2025

From my data of ~1.8M jobs, "Software Engineer" as a job title in postings are down ~25% and "Senior Software Engineer" jobs are down ~10% comparing the first 2 months of 2026 with 2025

I'm building out some interesting features on my app (UnlistedJobs). Basically I have job seekers as paying customers (not employers) so I can do things with the job listings that others can't - like tag a job posting if it is 1 standard deviation below the avg salary or "this job has been posted N times already." These things wouldn't fly if the employer paid for the listing. I am wondering what else would be most helpful to job seekers that you couldn't do if the employer paid for the listing

I analyzed 1.8M job posting descriptions and looked for 63 descriptor words and phrases. Here are the top 10 for software engineers:

problem-solver (37.7%), collaborative (36.2%), passionate (31.2%), hands-on (29.2%), high-performing (19.7%), creative (17.3%), analytical (16.7%), proactive (16.3%), team-oriented (10.2%), decision-making (9.0%)

Post it to "Show HN"

https://jabbs.co/

1. Churn isn't as bad as I would have thought. There is also the use case for someone who has a job and wants to find the next one. But yea if someone wants to only use for 1 month, that is expected

2. I think companies (or the ATS's they use) might object to reposting a job listing on my site but I'm not doing that. You can search/sort/filter jobs but in order to view the job posting you will open a link back to the original job.

https://www.unlistedjobs.com/

26.3% of jobs found on company websites are not advertised on any job boards but found using this tool.

The app scrapes company's directly (24/7) and give the user:

- a head start over other job seekers

- access to thousands of jobs not listed anywhere else

- daily job filter emails with ability to be highly curated to reduce noise

https://www.unlistedjobs.com/

A different type of job search site that gathers job postings direct from company websites. About 1 in 4 jobs are not advertised on any sites (like LinkedIn or Indeed) but they are found going direct to company career page.

Side note: I found my last gig using this method so have now built it into a web app. It is a paid service but feel free to DM me for a free trial.

It is a paid platform so people remove it from "show hn" if not free. But about 30% of the listings are indeed unlisted so it is more than an aggregator. Sadly the feature to tag those listings used google serp results and they cracked down on bots a bit early this year. So it is a mix of unlisted (aka not on paid platforms like Indeed) and listed

Hey HackerNews! I've previously posted about a job search app I built that gathers listings from company websites. Well its now on ProductHunt for anyone to vote on.

There is a 7-day free trial there (promo code: PH1WEEK) if anyone is interested. Otherwise it is a small monthly fee (just FYI)

A few stats w the app: - 230k+ total open roles (mostly tech) - 60k remote jobs - 20-25% of those listings are "unlisted" (not found on other job boards)

Some job counts by dev skills: - 761 Rust - 780 Perl - 886 C++ - 1165 Kotlin - 1222 Golang - 1451 MongoDB - 1487 VueJS - 1533 Redis - 1632 Ruby - 1684 PHP - 2672 Angular - 2723 NodeJS - 3062 .NET - 3784 TypeScript - 7007 React - 7350 JavaScript - 8037 Java - 15130 Python - 17817 Scala

Not free (DM me for a free trial) but please try out https://www.unlistedjobs.com/ which gathers listings directly from company websites

The data is destroyed and no content from the web pages are reused or repurposed (each listing is merely a link + various tags that are created/associated upon viewing. My understanding is that public websites scraping is legal but repurposing their content might not be

I think you should always be searching for your next role. This will keep you informed about the market and know if the SWE skills are indeed shifting towards AI (this would become part of their job interviews, which I have not seen yet)

Ah thank you for the signup and feedback. I'm increasing the services to check for expired jobs now that a lot of companies seem to be expiring there job listings around this time of the year, so apologies for that (I want to show more transparency into how/when those job listings are indeed checked for expiry too)

You can currently 'archive' any job listing by clicking on the 3-dots dropdown menu on the right side of each job listing to remove it from your view (and it will disappear and only be accessible in the "Archived" section of your sidebar). But gotcha there should be a report method, thanks so much!

Thank you. Currently working on promo code functionality to give away to all the postings in Ask HN: Who wants to be hired? (December 2024) https://news.ycombinator.com/item?id=42297422

Oh wow that’s not good, thank you so much for pointing that out!

Yea that one is still a WIP (along with "R", "Rust", maybe a few more) because parsing out some of these tags is a little more difficult than others.

Ah I understand, that is an interesting use case. Currently the only searchable keywords are tagged on the job listings upon creation, but I can see if these keywords come up often enough to make tags for them. I'm not familiar with those Google products. Appreciate the comment

Thank you, appreciate this. I'm not certain about expanding this in that direction but that is certainly an interesting thought.

Yea I feel ya, I’m definitely not a good designer based on how long this stuff takes me. I think we are all not that bad but once you spend 6 hours on a component you really feel your dev skills could be much better used. Anyway here are some of the things I used on the landing page:

CSS - https://tailwindcss.com/ Components - https://tailwindui.com/ Logo - https://pixlr.com/editor/ Icons - https://heroicons.com/ - https://lucide.dev/icons/ Animations - https://www.framer.com/motion/ BG patterns - https://heropatterns.com/ - https://dev.to/bybydev/top-10-svg-pattern-generators-16h

Otherwise its just React

The search bar is for job title only at the moment. Under the "Tech" filter you can find "C++" and add it as a filter

This is great feedback, I'm very appreciative!

Yea there is a much better version of the search bar soon-to-deploy (which accounts for aliases like QA -> Quality Assurance) and it will match by word rather than the entire phrase (currently "software engineer" will not query "software test engineer"). Appreciate the callout here

You can find a toggle switch for "has salary" under the "other" filters which will show only those w salary, but good call perhaps that should be part of a sort feature (beyond just date)

The filters do need more love for sure. I like your examples for various metrics displayed in the UI. I did think it would be cool to have a Github-like array of squares that represent units of time with colors that show how it has been changing over time, would have to figure out how heavy of calculations those would be in real-time but I really like your idea here. Or a line chart might be better.

Many thanks for all the input!

Thank you :) I appreciate the request + feedback. I have a story in the backlog to add location-specific links to the landing page but I really like your idea of having a map (heatmap or something) to show densities of jobs.

So the inclusive vs exclusive filtering is something that I struggle to perfect here. I'm tempted to throw both in the UI (since its ready to go on the backend) but its hard to explain to users. One thing you can do that is not so obvious is add a tag for "Canada" but click on the tag again which will put a line through Canada and exclude that location from your filter (still need to have helpers to show users how to do that). The 'remote' tag is probably the toughest one to parse of a job listing because it might appear anywhere within the text, so there is some inaccuracies for sure but its improving I hope!

Ah I could probably add filters for company locations specifically too (so you can filter US companies), that's an interesting use case too.

Thanks for the compliment too, it has been really fun to build

Ohhh yea I run into this memory issue very quickly when scraping (especially if you have a large URL dataset then it will inevitably find a website with a giant bit of markup). So I have to set timeouts and blacklist timely requests but also completely reset the (headless) browser on 2-3 requests (which is overkill but I am restricted on memory for those workers). Feel free to drop me an email sometime (should be on my HN profile)

Selenium, although I'm using a wrapper library that uses it. I only query each company every few days or so which probably helps to not get banned IP-wise but also rotate them. But many of the company job links are through external sources too (lever, greenhouse, etc.) which don't seem to mind

The company data was gathered online for a long time until I found https://www.thecompaniesapi.com/ (which now is the source for much of that data)

Oh that's a really interesting idea. Yea I dislike the idea of charging the job seeker but have not found a good way to monetize companies (not that they even know about me anyway)

https://www.unlistedjobs.com/

Scraper of job listings directly from company websites. I found my last day job by using a scraper that visits company websites in search of job listings. Now I've turned it into an app for others to use and access jobs that are posted on company websites (rather than paid employer ads on Indeed or wherever). This gives the job searcher an advantage to find jobs not listed on job search sites and show the company you have taken time/interest to visit their site.

This is a repost of https://news.ycombinator.com/item?id=42350351 which was flagged (hopefully those reasons have been addressed) and would really like to continue the convo with those interested.

The job search view (URL linked) can be done without signup so hoping feedback can be given and you don't have to signup if you don't want to. You can also delete your account if you get cold feet (all data is indeed deleted).

This is a scraper (and UI) of job listing from ~200k company websites. Full transparency the service is $15.50/mo but you can always use the job viewer to see if there are relevant jobs in your field without signup.

Would love feedback and input. I built this out of frustration in the current job search tools. Employers don't pay for listings which is why I ask the job seeker.

Thanks and look forward to any discussion :)