Rethinking job information delivery

Why are there so many clicks to apply for a job online? Every day, I repeat some version of this byzantine maze:

Open a job board.
Click a job posting.
Click apply.
Re-select the same job posting from the company’s own career listing page.
Click apply for the second time.
Travel to a third domain to fill out an application form.
Click an “Autofill form from resume button”.
Become very disappointed, clear the form, fill it out manually.
Click apply for a third time.
Success?

And rightly so, you might be thinking:

It is 2025 John, just use an agentic solution to do all of this for you and more?!

which is both a fair and accurate!

But, I think it misses the point - processes for humans shouldn’t be that complicated to begin with. I fear we risk incentivizing complexity when we naively view AI as a smoothing function for irregularity or difficult technology pain points. So if we peel back a few layers and want to get at a better system from a root cause perspective, what might that look like?

That’s why I created jobs.json: a proposed minimal, well-known file specification that allows companies and organizations to uniformly communicate open roles and career information, effectively and fairly.

Check out the full spec

Instead of every company designing its own careers portal with a different structure, jobs.json provides a consistent schema that can live right at the root of a company’s website (think google.com/jobs.json). It’s machine-readable, human-understandable, and built to make discovery effortless.

{
  "version": "0.1",
  "company": {
    "name": "Your Company Name",
    "url": "https://yourcompany.com",
    "jobsUrl": "https://yourcompany.com/careers"
 },
  "jobs": []
}

Imagine:

• Job boards could automatically sync listings from verified company domains.

• Candidates could search across employers directly from their own tools.

• Companies could reduce vendor lock-in and surface open roles without middlemen.

Most importantly, the entire hiring discovery process could be simpler, more transparent, and less clicky (which feels like a small, but meaningful, win).

The web already has conventions like robots.txt and security.txt, well-established specs that help systems talk to each other safely and (ideally) predictably.

But it also has humans.txt, a gentler, more humanistic idea: a place to celebrate the people behind the site. It’s not a standard, but a suggested statement, a reminder that the web isn’t just machines exchanging data, it’s people building connections. While it isn’t an IETF standard, it hasn’t stopped the idea from spreading over the last decade.

jobs.json aspires to live in that same spirit. It’s about giving people, not platforms, a simpler way to find each other and see what we can make together. The spec is completely open source with the hopes it can grow and evolve and bring values to both companies and future employees.

And if you’re hiring reliability/devops engineers (or just cool people in general), reach out to me (linkedin.com/in/johnhmccall), no jobs.json required!

Share