Where AI roles appear first
Almost every company above a few dozen people runs hiring through an applicant tracking system, and that system publishes a public board. Greenhouse boards live at boards.greenhouse.io, Ashby boards at jobs.ashbyhq.com, Lever boards at jobs.lever.co, and Workday and BambooHR each have their own hosted pages. When a hiring manager gets a requisition approved, the recruiter opens it in the ATS and it is live on that board within minutes. Nothing else has to happen for it to exist.
The volume backs this up. On TixelJobs, 19,557 of the 50,447 live listings come from Greenhouse boards, 9,702 from Ashby and 6,937 from Lever. LinkedIn contributes 5,895 and Indeed 5,262. The company boards are not a niche source; they are where most of the supply sits.
Why the aggregators lag
LinkedIn and Indeed get a job in one of two ways: the company pays to post or syndicate it, or a crawler finds it. Paid posting is a deliberate step that someone has to take, and at many companies it happens days after the requisition opens, if it happens at all. Crawling is faster but not instant, and a crawler that finds a posting still has to classify and publish it.
The delay matters because applicant volume is front-loaded. A recruiter reviewing a role in its first two days reads applications one at a time. Once the post is syndicated and the pile grows, screening becomes a search for reasons to reject. You want to be in the first group, and that means seeing the role when it is still only on the board.
Build a watchlist of company boards
Start with 30 to 50 companies you would actually join. For each one, find the board URL rather than the marketing careers page, because the board is what updates. A quick way is to open the careers page, click any job, and look at the URL of the application form; it will name the ATS.
Put the URLs in one document with the date you last checked. If you can write a few lines of code, a script that fetches the JSON endpoints and diffs the job IDs against yesterday's run is a two-hour project and removes most of the manual work.
- Greenhouse: boards.greenhouse.io/<company>. The board has keyword and department filters, and boards-api.greenhouse.io/v1/boards/<company>/jobs returns the same list as JSON.
- Ashby: jobs.ashbyhq.com/<company>. Filters by department, location and remote status are built in.
- Lever: jobs.lever.co/<company>. api.lever.co/v0/postings/<company> returns JSON if you want to script a check.
- Workday and BambooHR: the URL is company-specific. Bookmark the search page with your filters already applied.
A daily routine that takes fifteen minutes
The routine only works if it is short enough to keep doing. Fifteen minutes at the same time each day is the target, and these are the steps that fit in it:
- Check the watchlist at the same time every day, ideally in the morning in the company's time zone. Most recruiters publish during their working hours.
- Filter by the two or three titles you are targeting. Do not read every listing; a title and location scan is enough to spot what is new.
- Apply the same day to anything that fits. Tailoring a resume for a strong match takes 30 minutes, and waiting for a better moment costs you the early slot.
- Log every application with the date, the source and the board URL, so you can follow up after ten working days without guessing.
- Once a week, prune companies that have posted nothing relevant for a month and add two new ones.
Where TixelJobs fits, honestly
TixelJobs does the watchlist work at a larger scale: it reads AI and machine learning roles from more than 1,000 company career pages on Greenhouse, Ashby, Lever, Workday and BambooHR, plus Indeed, LinkedIn, Google Jobs, Glassdoor, Y Combinator and RemoteOK, twice a day, filters them for AI relevance and reviews them before they go live. That is the same thing your script would do, with more boards and a human check.
It is one way to do this, not the only way. If your target list is ten companies, a set of bookmarks is enough and you should not pay anyone. If your list is a few hundred, or you do not know which companies are hiring for your specialty, an aggregator that starts from the boards saves the hours you would otherwise spend building and maintaining the list.
Where to look next
The ML engineer jobs hub on TixelJobs is the largest category with 12,399 live roles, and the research scientist jobs hub, the NLP jobs hub and the MLOps jobs hub cover the other titles most people watch for. The remote AI jobs page lists the 26,067 listings that are remote, and the full search lets you combine a title, a country and a level. Browsing is free. Membership is $9 a year or $4.90 every three months and unlocks the apply link and the full description on every job. Cancel any time from your billing page.