HOW IT WORKS
How JBHNTR searches the job market for you
JBHNTR turns your career material and preferences into a cross-source market scan, then gives you a shortlist you can understand rather than a feed you have to filter again.
Build the search profile
Upload at least one CV, choose where and how you can work, select the levels and types of role you would genuinely consider, and describe what you want in your own words. The narrative matters because titles alone rarely capture the work.
The profile is persistent. You do not rebuild the same filters every time you return. Additional CVs, cover letters, target companies and feedback add context, but they should improve an already useful profile rather than block first value.
Scan across the market
A scan reads the shared job corpus built from broad feeds, niche boards and employer career pages. Hard constraints remove impossible roles first. Semantic retrieval then looks for work that is similar in meaning even when the title is different.
JBHNTR does not promise to cover every listing on the internet. Source availability changes, employers use different systems and some markets are better covered than others.
Hard constraints before judgement
Location, work mode, contract type and other explicit constraints remove jobs that cannot work before expensive judgement begins. A semantically relevant role in the wrong country is still the wrong result, and filtering first is also what makes the judgement step affordable to run at corpus scale.
Meaning before title
Job titles are inconsistent between employers. Semantic retrieval compares the meaning of the posting with the profile and the stated objective, so equivalent and adjacent roles surface without treating every shared keyword as relevance.
Two-way fit
The strongest candidate roles are assessed twice:
- Fits what you want: whether the role, company, location and work match your stated goals.
- You fit their requirements: whether your background plausibly supports the main requirements in the posting.
A role may score well on one side and poorly on the other. Showing that trade-off is part of the product, not an error to hide. Keep both dimensions visible so a reader can decide whether a stretch is worth taking.
Reasons, not just a score
Every scored role includes a tier, two-way fit, the strongest reasons to consider it and the points to check. “Watch out” is not a disclaimer appended to a recommendation; it is evidence that the system is willing to say no.
Match scores are prioritisation tools. They are not hiring probabilities and they do not replace reading the original posting.
Act, and inform the next scan
Open the original posting, save the role, mark it as applied, dismiss it or rate whether it was worth showing. For a selected role, JBHNTR can prepare an editable first draft of a tailored CV or cover letter. You decide what is true, what sounds like you and what gets submitted.
Ratings, notes, saves, dismissals and application states provide context for future ranking. A fixed amount of feedback is not a guarantee of improvement.
Limits
A model can misread a posting, miss unstated context or overvalue language in a CV. Scores are not hiring probabilities. Verify the original role and correct every generated claim before you send an application.