|
aBagorn posted:Huh, whoa that you mentioned this. The team Im going to be building is going to be working with a substantial amount of med data. Basically building our model (and then subsequently apis and apps) for external users and some of the ML teams inside our division. That's awesome. Probably too early to say but do you know what tools you're going to use?
|
#
?
Apr 17, 2018 21:36
|
|
|
|
| # ? Jun 8, 2024 08:44 |
|
|
Good Will Hrunting posted:That's awesome. Probably too early to say but do you know what tools you're going to use? Not yet!
|
|
#
?
Apr 18, 2018 02:10
|
|
|
BurntCornMuffin posted:Detect sanity, presence of data when expected, quantity of data. Ensure that via logging or metadata, you can trace the life cycle of your data for diagnosis. Have some sort of dashboard that you can show your consumers, and ensure your support people get alerts and act on them within a reasonable SLA and issue public statements for anything that would affect anyone in your orbit. This is great - sounds like you guys have it figured out. Any blogs/resources about this/scaling into big data and choosing the right org structure, SLAs? Or is this one of those cases where you just have to be on the grapevine. ultrafilter posted:There are limits on what you can do on a single machine, so you will eventually have to use the cloud. R has interfaces for TensorFlow, Hadoop, Spark, and other distributed computing platforms, so that's not as big a jump as it would be otherwise. I'll be sure to flip through these resources!
|
|
#
?
Apr 18, 2018 04:23
|
|
|
Tezzeract posted:This is great - sounds like you guys have it figured out. Any blogs/resources about this/scaling into big data and choosing the right org structure, SLAs? Or is this one of those cases where you just have to be on the grapevine. I'd at least skim the SRE book.
|
|
#
?
Apr 18, 2018 06:58
|
|
|
All I that I stated came from being on the grapevine, and running into these problems. I actually don't feel like much of an expert, there's a lot I don't know. At the same time, being able to write that definitely makes me feel like maybe I'm better than I give myself credit for. Maybe I should start blogging and writing. Perhaps it'll help make me more compelling as a potential hire, too, especially since I have a few interests that are outside of what I've proven to be capable of in my career, would love to pivot my career towards, and I need something to prove some skill in those interests.
|
|
#
?
Apr 18, 2018 15:30
|
|
|
BurntCornMuffin posted:All I that I stated came from being on the grapevine, and running into these problems. I actually don't feel like much of an expert, there's a lot I don't know. At the same time, being able to write that definitely makes me feel like maybe I'm better than I give myself credit for. I'd click your ads. The internet is still pretty barren in regards to consolidated information about how to initiate this stuff in the real world. Sounds like a good opportunity to make some cash if you can get the word out.
|
|
#
?
Apr 18, 2018 16:22
|
|
|
BurntCornMuffin posted:All I that I stated came from being on the grapevine, and running into these problems. I actually don't feel like much of an expert, there's a lot I don't know. At the same time, being able to write that definitely makes me feel like maybe I'm better than I give myself credit for. Almost certainly. It's super easy to discount the things you know when faced with constant reminders of everything you don't. I mean if you program at all you spend all day figuring out stuff you don't know, if you know it, you put it in and move on to something else without thinking about it.
|
|
#
?
Apr 19, 2018 00:11
|
|
|
Have any of you implemented some sort of design review process for your engineering org? I'm at a place where the team to team quality on new things can be hit or miss, and I'd like to help people grow, implement something sane, and get a better product out of it. My initial thought was to have us make a design doc just prior to breaking work up into stories, schedule a review meeting, and then gently nudge people to come with questions prepped. Unfortunately, I've never seen a thing like this implemented before. What would be things to watch out for? Is this even a sane approach? Does something like this even work?
|
|
#
?
Apr 20, 2018 08:28
|
|
|
Visual design or architectural design?
|
|
#
?
Apr 20, 2018 08:44
|
|
|
Apologies, architectural.
|
|
#
?
Apr 20, 2018 08:46
|
|
|
I've seen it done in a few places, almost always in the format of some document either requesting input for potential solutions or review of a proposed solution. I've seen them variously called a Request for Information, a Request for Discussion, and Request for Comment, a Spec Doc, etc etc. They work pretty well as an exercise for getting everyone on the same page and making sure people feel like they've had input in decisions. The biggest thing to look out for is bikeshedding, you will get a LOT of it, and you need to have someone with the authority to nip it in the bud and/or move it into external meetings, ideally the author of the document. Edit: I've also seen situations recently where they've *not* been used which resulted in 3 months of wasted development effort when the newly formed architecture team vetoed it.
|
|
#
?
Apr 20, 2018 09:00
|
|
|
What's the scope of the design we're talking about here? Are these projects that run for a year or is it more like two weeks?
|
|
#
?
Apr 20, 2018 09:46
|
|
|
Projects vary, they generally run a month or two.
|
|
#
?
Apr 20, 2018 14:28
|
|
|
2nd Rate Poster posted:Have any of you implemented some sort of design review process for your engineering org? We're working on it now. Basically the idea is a day or two after a story is put into in progress the engineers working on it present their approach ad hoc just after standup. Maybe show some code but mostly we want to hit it pre-code when there's still time to change things easily. It's not hugely successful yet but mostly because velocity has been poo poo these last two sprints.
|
|
#
?
Apr 20, 2018 14:51
|
|
|
What's pretty common in my area is someone will write up a doc and send it around to collect comments. Depending on the doc quality there may be more or less discussion and back-and-forth required, but generally by the end everyone's on the same page for everything that really matters, and the doc can be used as guidance to break the project down into actionable tasks. Someone then creates a bug for each task, and links the bug to the relevant part of the doc. Writing the docs is then also a good artifact to have lying around for when performance reviews hit.
|
|
#
?
Apr 20, 2018 21:25
|
|
|
TooMuchAbstraction posted:Someone then creates a bug for each task Assume it's a bug before it's even coded out. I love it!
|
|
#
?
Apr 21, 2018 02:45
|
|
|
BurntCornMuffin posted:Assume it's a bug before it's even coded out. I love it! "Bug" is the term for the internal issue tracker. Bugs can be feature requests, which is what I usually file these as.
|
|
#
?
Apr 21, 2018 04:36
|
|
|
BurntCornMuffin posted:Assume it's a bug before it's even coded out. I love it! Test driven development taken to the extreme!
|
|
#
?
Apr 21, 2018 07:34
|
|
|
BurntCornMuffin posted:Assume it's a bug before it's even coded out. I love it! Based on my experience, this is a very reasonable thing to do.
|
|
#
?
Apr 21, 2018 15:46
|
|
|
What are good ways to ask about company culture and project planning in an interview? I've had a couple of interviews for team leader or senior engineer positions lately and have been getting back pretty generic answers so I think I'm not phrasing my questions effectively. A great list of questions was posted the last time we discussed how to look for red flags during an interview (or maybe it was in the agilefall/newbie thread) but I haven't been able to find it. LLSix fucked around with this message at 10:21 on Apr 22, 2018 |
|
#
?
Apr 22, 2018 10:18
|
|
|
LLSix posted:What are good ways to ask about company culture and project planning in an interview? I've had a couple of interviews for team leader or senior engineer positions lately and have been getting back pretty generic answers so I think I'm not phrasing my questions effectively. If the answers aren't specific probably your questions aren't specific, e.g. "can you tell me about the planning process for the last two projects you did?"
|
|
#
?
Apr 22, 2018 10:21
|
|
|
LLSix posted:What are good ways to ask about company culture and project planning in an interview? I've had a couple of interviews for team leader or senior engineer positions lately and have been getting back pretty generic answers so I think I'm not phrasing my questions effectively. Can't remember the source. quote:How long have each of you worked at this company?
|
|
#
?
Apr 22, 2018 10:54
|
|
|
Volguus posted:Now they're coming around so maybe in 10 years .NET will catch up to Java. Uh, I'm late on this, but this is a joke right?
|
|
#
?
Apr 22, 2018 11:09
|
|
|
a hot gujju bhabhi posted:Uh, I'm late on this, but this is a joke right? I think the point was that Java has a better overall ecosystem than .NET, but C# is a better language. I don't have enough insight into the Java ecosystem to make an argument one way or the other.
|
|
#
?
Apr 22, 2018 15:03
|
|
|
New Yorp New Yorp posted:I think the point was that Java has a better overall ecosystem than .NET, but C# is a better language. I don't have enough insight into the Java ecosystem to make an argument one way or the other. Me and my coworker wistfully wishing we were using Maven and not NuGet while enjoying many C# language features agree with this.
|
|
#
?
Apr 22, 2018 15:14
|
|
|
prisoner of waffles posted:Me and my coworker wistfully wishing we were using Maven and not NuGet while enjoying many C# language features agree with this. We just write C# grade plugins and everything gets pushed to artifadtory.
|
|
#
?
Apr 22, 2018 15:35
|
|
|
prisoner of waffles posted:Me and my coworker wistfully wishing we were using Maven and not NuGet while enjoying many C# language features agree with this. I'm not trying to talk about something I don't understand, I'll freely admit I have no idea what Java has to offer in this regard. But can you explain what .NET is lacking, specifically? I've not really had an issue in the last couple years or so. NuGet is sometimes flawed but ecosystem wise I feel like there's always been a library out there for whatever I've needed. Again though, it's just my experience, people working on different types of software probably have different experiences.
|
|
#
?
Apr 23, 2018 10:45
|
|
|
a hot gujju bhabhi posted:I'm not trying to talk about something I don't understand, I'll freely admit I have no idea what Java has to offer in this regard. But can you explain what .NET is lacking, specifically? I've not really had an issue in the last couple years or so. NuGet is sometimes flawed but ecosystem wise I feel like there's always been a library out there for whatever I've needed. Again though, it's just my experience, people working on different types of software probably have different experiences. I think most of our worst issues are due to the particular history of things at the company (i.e. self-inflicted) and not NuGet per se, but here goes: Maven plugins allow additional tasks to be brought into the system, NuGet is meant to be a closed-set-of-functionality tool (unless you open up the powershell escape hatch, I guess). In our case, this means we have Python and Gulp scripts of dubious quality running our NuGet instead of community-tested plugins doing special purpose tasks within Maven. Is it accurate to say that Maven actually defined a declarative model for Java project building, including dependencies? Whereas NuGet inherits its project model from visual studio and bolted on two different formats for dependencies before putting them in the visual studio project itself... and putting dependencies there is a preview, or something. So Maven wanted to do every build activity you can do from a single machine and is open both as source and to extension by plugins. NuGet is, in contrast, just for packaging and dependency management, has no ambitions about being able to fulfill every task you might want to throw at it (that's what powershell and Team Foundation BuildLord or something else with a terrible name is for, right?) and suffers a little from being a small fish in a churning pond of Microsoft Dev tools. C# vs Java library ecosystem fights are not for me.
|
|
#
?
Apr 23, 2018 14:58
|
|
|
prisoner of waffles posted:I think most of our worst issues are due to the particular history of things at the company (i.e. self-inflicted) and not NuGet per se, but here goes: It seems like an apples to oranges comparison. Maven is a build language with dependency management built in. NuGet is purely for dependency management. MSBuild is the Microsoft build language. And yeah, MSBuild kind of sucks, but that's kind of forgivable because the Visual Studio tooling handles most of it for you; I haven't had to seriously dig into an MSBuild script for a long time.
|
|
#
?
Apr 23, 2018 16:15
|
|
|
For me, NuGet vs. Maven is even more like "an apple in a bowl of fruit that I don't particularly like vs. a different bowl of fruit"* *The other point was that C# is (IMO) ahead of Java and some weird async, Task<T>, and reflection stuff I had to work with about 5 minutes ago was actually quite well explained in docs and on SnackOverflow as well as being Nice and Slick.
|
|
#
?
Apr 23, 2018 17:39
|
|
|
Maven does, however, have some very fucky ways of declaring and resolving test dependencies.
|
|
#
?
Apr 23, 2018 19:07
|
|
|
prisoner of waffles posted:I think most of our worst issues are due to the particular history of things at the company (i.e. self-inflicted) and not NuGet per se I think there's also a valid complaint right here. Anecdotally, there's a lot more NIH and homegrown libraries in .NET land, and it's way less likely for a non-MS-sanctioned library to see much adoption. prisoner of waffles posted:Team Foundation BuildLord
|
|
#
?
Apr 23, 2018 20:54
|
|
|
redleader posted:I think there's also a valid complaint right here. Anecdotally, there's a lot more NIH and homegrown libraries in .NET land, and it's way less likely for a non-MS-sanctioned library to see much adoption. I'm kind of weirded out how common Newtonsoft JSON is, but that's Microsoft's fault for lashing their first JSON support to some XML-transform bullshit. As in, "why the hell are there XML namespaces in my JSON property names?!", was the first thing I thought looking at the default serializer output, IIRC.
|
|
#
?
Apr 24, 2018 13:15
|
|
|
Munkeymon posted:I'm kind of weirded out how common Newtonsoft JSON is, but that's Microsoft's fault for lashing their first JSON support to some XML-transform bullshit. I had the same thoughts looking at the framework's JSON serialization back before JSON was commonly used. Instead of doing it right, Microsoft is probably responsible for thousands of wasted man hours and directly creating a few alcoholics due to dealing with binding-redirect/dependent-assembly hell with all the Nuget packages that take a dependency on Newtonsoft (most of them.)
|
|
#
?
Apr 24, 2018 14:14
|
|
|
After the backlash I got for all the prep work I was doing from here I took a step back and evaluated why and yeah, a lot of it felt like overcompensating for a lack of experience and severe anxiety from how lovely my current job is. I decreased the volume by quite a bit and tried to do a problem every few days and I feel pretty comfortable at this point. The only two things I really struggled with were "print all ways nodes may have been added to a BST" and "queens on a chessboard". Most other problems seem trivial at this point, in a comfortable environment at least. On the flip-side, we've wrapped up our first iteration of our project so I have a bunch of time off next week (lol first non-weekend days off in 4 months) and I would like to spend some time exploring some of the more challenging areas that are interesting to me that I haven't touched at all really. My question is: where do I even start? People have mentioned tree balancing problems, red-black, AVL trees, tries, but I'm not sure where to begin. Do I just jump in and think about solving the problems myself?
|
|
#
?
Apr 24, 2018 15:17
|
|
|
No judgment on you, GWH, and kudos on finding an understanding of your anxiety, but man "CS algorithms as shibboleth showing your worthiness to join a big tech company" seems like such a waste. That said, why not find some online collection of problems with a judge and do some of them (or keep doing them if that's what you started)? Deep dives into certain CS concepts will pay off best if you literally get asked about the concept, a good set of problems with a grader will stretch your skills and build your confidence with feedback.
|
|
#
?
Apr 24, 2018 15:40
|
|
|
http://open.kattis.com/ This site has a ton of problems and a ranking to boot!
|
|
#
?
Apr 24, 2018 15:46
|
|
|
Why not, I don't know, actually take a break? I know the US is some kind of hellscape when it comes to having a reasonable amount of holidays, but try doing something non-computer related. I'm serious. Go for a walk in the country or whatever, it'll be good for you. If you really, really, must do something that you think will help you work-wise, try something that isn't anything to do with your day job. Go through the exercises on http://www.4clojure.com or read https://www.scottaaronson.com/democritus/
|
|
#
?
Apr 24, 2018 15:51
|
|
|
Naar posted:Why not, I don't know, actually take a break? I know the US is some kind of hellscape when it comes to having a reasonable amount of holidays, but try doing something non-computer related. I'm serious. Go for a walk in the country or whatever, it'll be good for you. I'd love to take a break but I'd also love to get out of my job even more and plan to take time off between jobs even. Work is really grinding me down with the politics and games and completely absurd schedule (manager and PM on my team will literally not get in until noon+ and ping me at like 3am, I won't wake up of course but it's still frustrating to see that first thing waking up). If I was making progress in terms of my skills, this would be a lot easier to swallow. But I get nothing out of this save for a paycheck. Keetron posted:http://open.kattis.com/ Thanks, this looks cool! prisoner of waffles posted:No judgment on you, GWH, and kudos on finding an understanding of your anxiety, but man "CS algorithms as shibboleth showing your worthiness to join a big tech company" seems like such a waste. You're probably about 60-70% correct. I don't even want to work at a big company, but during some of my last interviews tough algo stuff got me bounced and it's one thing I can actually work to improve on myself in a specific and measurable way. I could build a pet project I guess, or contribute to some open source stuff too, but at the end of the day these types of problems are still the most common thing I've seen.
|
|
#
?
Apr 24, 2018 16:02
|
|
|
|
| # ? Jun 8, 2024 08:44 |
|
|
I guess the real takeaway is: how do you mask lovely job experience where you're not given much of an opportunity to progress?
|
|
#
?
Apr 24, 2018 16:04
|
|