Developer builds an AI agent that applies to HH.ru job listings automatically

A developer has published a detailed write-up on Habr describing how they built an autonomous AI agent for job hunting on HH.ru. The bot searches for listings, weeds out irrelevant ones with a local neural network, writes a personalized cover letter for each position and sends reports to Telegram, with no spending on paid APIs.

Why job hunting turned into a war with robots

The author describes the IT job market as ruthless to newcomers: to get even one reply you need to apply to hundreds of listings, and what usually comes back is automated rejections from HR systems. Algorithms now work on both sides of the process: employers filter résumés automatically, and a job seeker has to write a unique cover letter for each listing to get past those filters. The author decided to offload exactly this routine to their own agent, instead of manually rewriting the same text for hundreds of different job descriptions.

What the homemade AI agent can do

The idea behind the agent is simple: it works autonomously and fully covers the routine cycle of applying to jobs while the user gets on with their own business.

• Searches for listings on HH.ru on its own
• Filters out irrelevant and "junk" listings using a local neural network
• Generates a personalized cover letter for each listing
• Sends reports on the listings found and applications submitted to Telegram
• Runs entirely for free, with no paid APIs and no extra setup

Why a local neural network

The author set a hard condition from the start: the script must be free and self-contained, with no fiddling around paid APIs. So the filtering of listings, the highest-volume and most routine part of the process, is done by a local model that runs right on the user's machine and requires neither a subscription nor payment for tokens. This makes the agent fully independent of external services and their limits: no matter how many listings need processing in a day, the cost of running it does not change and does not grow with the intensity of the search.

The cover letters, meanwhile, stay personalized for each specific listing rather than being mass-mailed as one template. By the author's reasoning, this is what should improve the chances of getting past HR systems' automatic filters, which easily spot identical text sent to dozens of employers in a row. In this setup Telegram serves as the single point of contact: the agent works in the background and sends reports on its activity, without requiring the user to constantly supervise the process or launch each stage manually.

How this agent differs from mass-mailing applications

The key difference in the approach is that it rejects simple application spam. Many job seekers tackle the sheer number of listings by mass-mailing the same résumé and cover letter to everyone, but that approach is the quickest to trip employers' automatic filters. The author took a different path: the agent first weeds out irrelevant listings with a local neural network and only then generates a letter aimed at the specific job description. As a result, the number of applications may be lower than with mass mailing, but each of them, in theory, looks more meaningful to the recruiter or HR bot reading it.

What this means

The homemade AI agent for HH.ru is a characteristic example of job seekers in an overheated IT job market picking up automation tools themselves: since employers' algorithms screen résumés en masse, some candidates are answering in kind and building their own bots for applications and correspondence, spending neither extra money nor extra time on a routine that used to be done by hand.