[RicoElectrico]

#pracait #programista15k

https://news.ycombinator.com/item?id=42533623

I'm a software manager that has been doing some form of interviewing/hiring for 13 years.

I did two rounds of hiring software engineers last year, one in spring that seemed normal, and one in the fall that was was brutal. The fall hiring had a flood of applicants, and in retrospect, most seemed like AI was used in some way.

For the fall round, I suddenly had a higher percentage of applicants that qualified after resume screening and initial phone screen, but they all collapsed when I did a technical round. And failure rate on the technical was much much worse than usual.

We have a full-time recruiter working with us, and I'm not 100% sure what tools he used, but I switched to manually reviewing each resume, and given that it was 100s, it took a long time, but I still had my problem of great initial screen, terrible technical interview.

Then, I decided to throw out anyone who heavily mentioned AI, LLM, or data science. After all, with almost a thousand applicants, I needed to sort some how. (To be fair, our use case is more esoteric, we're not writing Javascript or parsers, so it's not as much of a time-saver.) Large chunks of applicants dropped and the process felt more normal.

I also switched to only on-site only. My initial technical screenings are still done remotely.