Developer builds an AI agent that automatically applies to jobs on HH.ru

A developer has published a detailed Habr write-up on how he built his own autonomous AI agent for job hunting on HH.ru. The bot searches for vacancies, uses a local model to filter out irrelevant listings, writes a personalized cover letter for each position, and sends reports to Telegram, all without spending money on paid APIs.

Why job hunting turned into a war with robots

The author describes the IT job market as ruthless toward newcomers. You have to apply to hundreds of openings just to get a single reply, and most of the time the only responses are automated rejections from HR systems. Algorithms now run on both sides of the process: employers screen résumés automatically, and to get past those filters, candidates have to write a unique cover letter for every opening. Rather than manually rewriting the same text for hundreds of different job descriptions, the author decided to offload that busywork to an agent of his own.

What the homemade AI agent can do

The concept is simple: the agent runs autonomously and handles the entire routine application cycle while the user gets on with other things.

• Searches for vacancies on HH.ru on its own
• Filters out irrelevant and "junk" listings using a local model
• Generates a personalized cover letter for each vacancy
• Sends reports on found vacancies and submitted applications to Telegram
• Runs completely free, with no paid APIs and no extra configuration

Why a local model specifically

From the start, the author set a hard requirement: the script had to be free and autonomous, without jumping through hoops to wire up paid APIs. So vacancy filtering, the highest-volume and most repetitive part of the process, is handled by a local model that runs directly on the user's machine and needs neither a subscription nor per-token payments. That makes the agent fully independent of external services and their rate limits: no matter how many vacancies it has to process in a day, the cost of running it stays the same and doesn't grow with search activity.

The cover letters, meanwhile, are still tailored to each specific vacancy rather than mass-mailed from a single template. In the author's view, that is exactly what should improve the odds of getting past automated HR filters, which easily spot identical text sent to dozens of employers in a row. Telegram serves as the only point of contact in this setup: the agent runs in the background and reports on its work, so the user doesn't have to constantly monitor the process or manually kick off each stage.

How this agent differs from mass-blasting applications

The key difference is that the approach rejects simple application spam. Many job seekers tackle the volume problem by blasting the same résumé and cover letter to every opening they can find, but that is precisely the approach that trips employers' automated filters fastest. The author took a different route: the agent first uses a local model to screen out irrelevant vacancies, and only then generates a letter aimed at the specific job description. As a result, the agent may send fewer applications than a mass mailing would, but each one theoretically looks more meaningful to the recruiter or HR bot reading it.

What it means

The homemade HH.ru agent is a telling example of how job seekers in an overheated IT labor market are picking up automation tools themselves. Since employers' algorithms now screen résumés at scale, some candidates are responding in kind, building their own bots to handle applications and correspondence without spending extra money, or extra time on the busywork they used to do by hand.