That works for startups, but most larger companies still won't hire you without an interview, your old collegaues can only refer you to hiring.
Very often that's where I screw up. I code computers, not goddamn CoderPads. I run my shit to know if it works, not pretend I'm an interpreter and go through line by line. I cut and paste boilerplate and edit it, not write the boilerplate.
Some dude in an interview at a very well known AI company was staring at me on a video call and grilling me about the max value of a loss function whose formula was written in an image, and I didn't have a whiteboard to do the math on. TF do I care? I usually minimize losses, not maximize them.
Tech folks often complain about the interview experience but a big part of that experience is created by tech folks. Stuff like what you went through is really to blame in tech culture imo
Yeah exactly. I mean really, I have 15+ years of experience and they're grilling me on a correct implementation of quicksort and are going to fail me for an implementation that doesn't terminate when I had nothing to test it on?
I WILL stumble because it's not the kind of thing I do. That's the kind of thing a fresh college grad would have memorized and can vomit it to you, maybe without even understanding it. Maybe I'll code you a basic WebRTC conference call app, or train a simple image classifier, or outline an architecture of a massive system for you?
I'd be wary of almost any startup that hired without any form of interviewing whatsoever. The big difference is that you walk into the interview as a default-yes rather than a default-no, ie: they're on the lookout for any red flags that would cause them to have serious reservations but if no flags pop up, you're assumed in.
At big companies, there's more of a formal process but the key is get the attention of the hiring manager and make sure that hiring manager has enough juice to quietly backchannel to whatever process to treat this as a default-yes conversation. A good sign of how effective your manager is at a Bigcorp is how well they know how to work their internal system to get the right people onto their team.
Very often that's where I screw up. I code computers, not goddamn CoderPads. I run my shit to know if it works, not pretend I'm an interpreter and go through line by line. I cut and paste boilerplate and edit it, not write the boilerplate.
Some dude in an interview at a very well known AI company was staring at me on a video call and grilling me about the max value of a loss function whose formula was written in an image, and I didn't have a whiteboard to do the math on. TF do I care? I usually minimize losses, not maximize them.