4 ms·
This was an interesting observation: > What I do know, however, is that for every 1-hour interview where I evaluated if someone knew their data structures, I c
by ChrisMarshallNY 3y ago
This was an interesting observation:
> What I do know, however, is that for every 1-hour interview where I evaluated if someone knew their data structures, I could have just taught them.
I don't really hear much about training. I doubt it's because we don't do it, but maybe it's not an interesting topic for discussion.
- angarg12 3y agoSounds a bit delusional to me that you can just "teach someone data structures in 1 hour". Also the role specifics matter here. It might be ok to bring in a junior who needs lots of mentoring onboard. For a tech lead who is going to be driving the direction of a team, I'd expect them to be up to speed and autonomous very quickly.
- spankalee 3y agoIt's not "teach someone data structures in 1 hour", but teach them the approach needed for that problem in 1 hour. And it's a pretty insightful comment, because being able to be taught a practically useful algorithm, and use it on a question, in just an hour is a sign of a good candidate. Interviews should often be run just like that.
- Paul-Craft 3y agoI don't think "I can teach someone data structures in 1 hour" is a good faith interpretation of what OP meant. I would expect it's more like "I could teach someone about a data structure that could be used to solve this problem efficiently and simply in 1 hour" is closer to the mark.
- GoToRO 3y agoTraining is not billable. Managers decided they only like billable hours. Kind of like when you decide that you only like winning the race and you hate training.
- mrmetanoia 3y agoI dropped out of school (English literature) because I needed to make money and pay bills due to a life change, and i applied at a place that does internet stuff to do tech support because computers were a big part of my hobbies. I moved up over the years learning as I went and now I've occupied a few 'engineering' roles. I used to feel somewhat embarrassed to hold an engineering role with no engineering degree. But I stopped worrying the more I worked with software engineers. They were just people, some sucked, some were great, their education almost never had anything to do with the bracket they fell into. The worst one I've ever met has a PhD. The best one I've met also has a PhD. Then I found more people in these positions that have degrees or no degrees, and the people that work well both with degrees and without had something in common. They can more or less be taught to do anything, they can extrapolate and apply knowledge outside of the specific example initially given, and they understand the big picture/desired end results. I hate to say it so crass, but they are good at identifying the trivial bullshit and addressing it or cutting through it rather than sitting in it. They "Get it." From the ideal to what the business demands, they just get it. Training competent people is a boon. I don't know how you seek out interested/curious competency and then train it, but if we can figure that out, it'd be cool. I'm really tired of working with people who have degrees and jobs because their mothers told them it pays well. Computer Scientists with no interest or curiosity about computers.
- majormajor 3y agoTraining and teaching is hard. "One hour of interviewing could be avoided with one hour of training" is wildly optimistic. It also ignores the breadth of knowledge you likely want a candidate to have (and are just trying to sample at through a short few interviews). How many hours of training are you willing to sign up for? How confident are you that the knowledge will "stick" for any given candidate? Are you going to fall back to GPA and school prestige to measure how well they've retained knowledge in the past? That seems like a step in the wrong direction.
- xandrius 3y agoYeah but a data structure is definitely some stuff one could grok in 1h.
- darrenkopp 3y agoI don't think it's wildly optimistic, but perhaps we are thinking about it in different ways. I don't think I could teach someone how to implement each data structure in an hour, but I could easily go over maps/queues/stacks/hashtables and tell you when to use each in an hour. I know this because I do that very thing in less than an hour in code reviews. I do agree that the "stickiness" is iffy, but it usually sticks pretty well. You then have a second problem that can be described as "when you have a hammer, everything is a nail" as they use their new fancy hashtable everywhere.