Hacker News Reader

Comment by Nbox9 | Hacker News Reader

I’d be happy if some organizations started accepting a credential in lue of a coding interview, and just did a design/architecture and cultural interview.

View on HN

People here are suggesting credentialing as if it will remove interview requirements rather than simply add a new hoop to jump through. I do not believe required credentialing solves the interviewing problem at all. You see it today with the certifications that already exist in the market. They aren't worth the paper they're printed on. Anyone can trivially study and pass the certification tests and still never have implemented a system, nor even have basic programming ability.

A real credential with teeth would likely eliminate plenty of the "Does this person know anything at all" element of current tech interviews. I've interviewed people who didn't even know what a for loop was or the difference between the stack and the heap. If there was a piece of paper which said that the candidate knows at least the bare minimum about software development, you could at least start the interview at a more advanced (or domain-specific) level.

I doubt hospitals interviewing senior doctors need to ask them basic anatomy questions, or law firms needing to ask candidates the difference between tort and criminal law. Tech could benefit from this minimal minimal bar.

Just turn the Leetcode-style data structures/algorithms segment into the basis of this credential, so it at least doesn't have to be repeated ad nauseam each time an applicant interviews with a different company, despite having the least relevance.

Triplebyte was trying to do this a while ago, weren't they? My last observations from a year and change ago suggested they were running into what might have been signaling equilibrium issues / risk aversion, but I'd be curious if there's other perspectives around.

That's what Triplebyte is attempting to do, yep. But it does seem like a considerable challenge and they're still figuring out their product. Also feels like for an industry-scale problem like this, it would require a consortium of the major employers (not just FAANG but large companies from Microsoft to Intel to Oracle and beyond) to hash out some sort of standard, not to mention a body representing the engineers (if not a SWE union, at least something like the IEEE/ACM) and the academic institutions that provide the education.

One wonders what the history of how the credentials in medicine or law were forged.

Adding a new hoop is the desired effect, and it does solve at least one problem. If you're at a tech company or some other org that interviews well, you might not have seen it yet.

Many devs are not good at their jobs. They mean well, but they can't solve basic problems without looking at stack overflow. And by "basic", I don't mean leetcode, I mean iterating over a collection.

> Anyone can trivially study and pass

Yes! At that point, I would know that the guy sitting next to me did at least some amount of studying of the fundamentals.

> I’d be happy if some organizations started accepting a credential in lue of a coding interview

Sure, but what credential exists that both is reliable enough to support that use and would remain reliable enough once it became used that way, even in a specific narrow subfield of development?

A majority of Silicon Valley-style tech is now using data structures/algorithm interviews, so it already is a de facto credential.

What is the interview process like for architects? Lawyers? Surgeons?

I can't speak for architects, but there is a very high bar for licensing for them. For lawyers, I had a friend at a top tier firm that interviewed with another and had to do a full doc review- essentially doing several hours of actual work. They were moving to a small boutique firm outside of "big law" so this may have been somewhat outside the norm.

Reply on news.ycombinator.com