Automated resume analysis systems and artificial intelligence often filter out experienced candidates due to absurd technical formalities. This was proven by Android developer Artem Pelevin, who decided to check his own CV through popular job aggregators after reading materials about the "human resource shortage" in the market.
Automated resume analysis systems and artificial intelligence often filter out experienced candidates due to absurd technical formalities. This was proven by Android developer Artem Pelevin, who decided to check his own CV through popular job aggregators after reading materials about the "human resource shortage" in the market.
According to an IT specialist with 4 years of experience in Android/Mobile development, the systems estimated his relevance for profile vacancies at only 15–60%.
“So what result did I get from systems that automatically analyze skills? Simply terrible. Terrible to the point of absurdity. In 4 years of working in Android/Mobile development, my relevance is 15–60%… What? And here begins the hell of automated resume analysis systems,” Pelevin said on LinkedIn.
The developer noted that the algorithms of aggregators (in particular, thejob.tech) are unable to connect basic synonyms and abbreviations that are obvious to any tech specialist or recruiter. In particular, the system "did not see" Android skills in a specialist in the position of Android Developer, defined KMP as an irrelevant technology for Kotlin Multiplatform, and also broke the connections between REST API and REST, Git and GitHub, SQL and SQLite.
In addition, the developer drew attention to the unstable operation of artificial intelligence models (LLM) used for initial screening.
“I don’t know why in some vacancies Java is highlighted in green, while in others it is indicated that this skill is not in the resume… Most likely, the system is simply saving tokens and not analyzing the resume completely. The names of some technologies can be written differently, and your technical specialist will understand this. The recruiter can ask or Google it. But LLM will be eliminated due to a mismatch,” the specialist emphasized.
To prove his point, the developer provided a screenshot with the funniest algorithm glitches. It shows how the system counts Java knowledge in some vacancies, but “doesn’t see” it at all in others. In addition, the absence of the word “Android” in capital letters in the hard skills section led to the algorithm simply ignoring this skill, despite the clearly indicated position of Android Developer in the CV.
Artem Pelevin urged recruiters not to rely entirely on algorithms and to manually screen candidates, as calls about a "staff shortage" often arise precisely because of automatic filtering errors. He advised fellow developers to independently test their own CVs through such aggregators and adapt the wording to the requirements of robots.
