About Workhunters
Workhunters is an AI job search agent. It indexes openings directly from employer career pages, matches them against your profile, and then fills out and submits the applications for you. The job seeker chooses the jobs; the software does the data entry.
Problem
Openings are scattered
Most real openings live on an individual employer career page, not on a job board. Finding them means checking thousands of separate sites, each with its own layout and its own idea of what a job posting is. Aggregators paper over this with reposts, duplicates, and listings that were filled weeks ago.
Applying is repetitive data entry
Workday, Greenhouse, Lever, and the rest ask the same questions in a different order every time, often behind a new account, an email confirmation, and a verification code. A single application can take twenty minutes, almost none of it spent on anything a hiring manager will ever read.
The volume required is unrealistic
Response rates mean a serious search runs to hundreds of applications. Doing that by hand is a full-time job, which is exactly what the people doing it do not have. The work gets abandoned long before the search is over.
Who it is for
Active job seekers applying at volume
People running a real search who need to be in front of many employers, not five. Early-career candidates and recent graduates feel this hardest, because the volume required to get a first response is highest for them.
Candidates searching across borders
People relocating or open to another country, who need listings filtered by work authorization, visa sponsorship, and language requirements rather than by a single city name.
People with less time than the search demands
Anyone job hunting around a current job, caregiving, or study. Reviewing matches takes minutes. Submitting them takes none.
Team
George Petrovski
George has worked in sales, at recruitment agencies, and in software development roles. That is both sides of hiring: screening candidates, and being one. Workhunters comes out of the second, and out of first-hand experience of how much of a job search is spent on work that has nothing to do with whether you can do the job.
He founded Workhunters and built the platform end to end: the crawler fleet that indexes employer career pages, the AI extraction pipeline that turns those pages into structured listings, the matching engine that ranks them against a profile, and the browser agent that completes and submits the applications.
The project started in 2018 as a personal tool for tracking postings scattered across company career sites, and was rebuilt into a product as language models became good enough to read a job page the way a person does. It has been worked on continuously since.
The machine learning is his own work at every stage: assembling and labelling the training data, choosing and training the model, tuning it against held-out results, and running it in production under real load. Job postings are read by a large language model served on hardware Workhunters owns rather than rented from a third party. Incoming employer email is sorted by a classifier he trained on a hand-checked corpus of real messages, which files most of them in about a millisecond and passes only the uncertain ones to the larger model.
Owning that stack end to end is what makes the economics work. It is why a free plan is possible at all, and it keeps candidate data on infrastructure the company controls.
Product
Job discovery index
Live in productionA crawler fleet visits employer career pages on a daily cycle and an AI extraction pipeline reads each posting into a consistent format: title, location, seniority, pay, language requirements, and visa sponsorship, whatever shape the original page was in.
AI matching and recommendations
Live in productionYour resume and preferences are turned into a profile, and new listings are scored against it every day. Recommendations are free and cost nothing to review or dismiss.
Application agent
Live in productionA browser agent opens the employer portal, creates an account if one is required, fills in the form, uploads the resume, handles the email or SMS verification code, and submits. If it cannot finish, the credit is refunded automatically.
Application tracking and inbox
Live in productionEvery submission is tracked through queued, applied, interview, and rejection, and exportable to CSV. Each account gets a dedicated Workhunters email address, so employer replies arrive in one place instead of buried in a personal inbox.
Employer-funded applications
In developmentThe next step is letting employers fund applications from well-matched candidates, so that applying costs the job seeker nothing at all.
Workhunters is live and open to the public. The platform runs on Google Cloud with its own GPU inference hardware, applications are submitted against real employer portals every day, and paid plans are available alongside a free tier. The company is early: the product is in front of its first customers and is being improved on what those real applications reveal.