|
ohgodwhat posted:It gets less accurate as incomes increase because more of your income goes towards savings. My cost of living is $60k and moving to NYC, maybe it's $90k, a $30k increase rather than the $100k increase that metric would predict. So you're looking for a salary survey, then? https://www.oreilly.com/ideas/2017-software-development-salary-survey you can scroll allll the way down to the bottom under "The Model in Full" to see what modifiers to apply to yourself based on your skillset to adjust what you'd expect to make. Robert Half also publishes one that's more of a forecast based on what they see their army of consultants and full-time placements getting: https://www.roberthalf.com/salary-guide/technology It'll have similar sections where it'll tell you to add some money for certain skills or locations. Those are going to be more precise than COL adjustments, but that precision requires you know stuff like team size and what stack/language you're going to be working with.
|
# ? Oct 26, 2017 14:15 |
|
|
# ? Jun 5, 2024 04:44 |
|
Zaphod42 posted:At some places its no reason, but not necessarily. Some places do use 'software programmer' to mean basically code monkey / entry-level, where 'software engineer' is a step above where you're actually doing more high level design work. And then you've got 'software architect' above that, and yadda yadda. It's worth understanding the standard title ladder that most big tech companies use, and trying to keep your titles in line with that. If nothing else, it gives people evaluating your resume an easy way to get some context into your responsibilities when you're looking for another job.
|
# ? Oct 26, 2017 14:20 |
|
All I know is that people treat “engineers” and “architects” with more respect than “programmers”, “coders”, and “developers” in terms of perceived skill and seniority, compensation, and personal development. That’s why people are so sensitive to job titles. There’s a measure of social ladder-climbing that comes with those labels, and as someone who is already an abject minority I am remiss to be considered lesser or incapable. I call myself an engineer not just because I have a STEM degree and because it describes me well, but because I feel the need to defend my validity and agency. Basically, I have a massive inferiority complex and the question of titles and seniority feeds it. vv
|
# ? Oct 26, 2017 15:07 |
|
Pollyanna posted:Basically, I have a massive inferiority complex and the question of titles and seniority feeds it. vv This is so surprising to hear.
|
# ? Oct 26, 2017 15:14 |
|
Other than CodeNewbie, is there any programming related podcast I can listen to during the non-thinking portions of my regular workday that would be understandable for a person getting started?
|
# ? Oct 26, 2017 16:59 |
|
I really like The Changelog, and Hanselminutes occasionally has good guests.
|
# ? Oct 26, 2017 19:02 |
|
Munkeymon posted:So you're looking for a salary survey, then? https://www.oreilly.com/ideas/2017-software-development-salary-survey you can scroll allll the way down to the bottom under "The Model in Full" to see what modifiers to apply to yourself based on your skillset to adjust what you'd expect to make. That's exactly what I wanted! Thanks!
|
# ? Oct 27, 2017 02:06 |
|
Hellblazer187 posted:Is this the right thread for people who are not (yet) working as programmers to discuss learning to program? Is there a thread for that if not this one? Stuff like just talking about the challenges of it (both academically and outside of that) and comparing resources, sharing success, etc. Along those lines, I was hoping I could get some input on what I should do and focus on. I decided to switch careers last year and just completed a year-long comp sci cert at a university. Since then I've been self-learning machine learning stuff, as that's where my interests are right now. I've done some coursera and udacity, so I have a few projects in deep learning on my github from them. The problem is that, to me, deep learning seems insurmountable to realistically get a job anytime soon, especially with no industry experience and very little knowledge. There's just an insane amount to learn, and I don't know what practical skills I should develop. I have the projects, but I can't honestly tell you in depth what a CNN is and how it works for example. I'm also afraid that it's just the current dumb fad and next year they'll be hiring block chain fidget spinner engineers instead or whatever. Should I continue what I'm doing and essentially gamble that I can get an internship next year as a machine learning engineer? What should I be focusing on, then, specifically? Google's Brain internship is open for application until January--I know I won't get in, but part of me wants to try anyway. There also seems to be a lot of positions open everywhere in general, but I haven't tried yet. The reason I'm asking is because of a lot of personal doubt and poo poo I've been going through lately, but mostly I'm 30 and unemployed now. I'm not currently hurting for money, but I'd like to at least feel like I'm making things happen and get my head straight.
|
# ? Nov 1, 2017 06:36 |
|
AlphaKeny1 posted:Along those lines, I was hoping I could get some input on what I should do and focus on. I would consider setting your sights lower than machine learning. There are not a lot of (good) jobs as a junior developer in machine learning. If you want to get into development professionally, some aspect of web development is likely a lot easier to get your foot in the door, and then move from there as you find things that interest you.
|
# ? Nov 1, 2017 06:43 |
|
AlphaKeny1 posted:Along those lines, I was hoping I could get some input on what I should do and focus on. What's your background? I feel like if you're STEM, then ML is completely ok because so much of it is math/statistics and the other part of it is database work. Skills in math and general programming are as evergreen as it gets, so don't sweat it. If the field implodes and you really need money, you can always go back to your old career. Send out lots of applications, mock interview your brain out, network/get connected with people currently in the field to get an idea of what they're facing. As for stuff proven in industry, collaborative filtering (recommendation algorithms used in Amazon/Netflix), decision trees (used to tell statistical stories), regression, possibly SVMs. Neural networks are the new hotness - check out all the stuff from Geoffrey Hilton because he really popularized this stuff recently. Just understand the backpropagation on simple 1-2 layers and how to implement it and what success/fail cases look like for analysis. And no one really knows how Deep Learning works other than to throw GPU processing power and more layers at problems
|
# ? Nov 1, 2017 11:08 |
|
My background is completely unrelated to STEM. Getting a certificate in comp sci was me trying to open up opportunities and stop working dead end jobs. If I drop machine learning for now and try to get a junior dev position, do I just try going for anything and everything? Right now, I'd prefer not getting stuck in web and front-end long term. I guess my question is, how does the industry as a whole work in terms of career path in that regard? Like if I'm working in the bay area as a junior web dev, what are my chances of getting stuck in that field versus moving into stuff that I actually want to do? I think I'm hesitant because of my unfamiliarity with how the industry works in general. Also web presents its own problem, because I haven't touched javascript and its million frameworks.
|
# ? Nov 1, 2017 20:23 |
|
AlphaKeny1 posted:My background is completely unrelated to STEM. Getting a certificate in comp sci was me trying to open up opportunities and stop working dead end jobs. It’s pretty easy to switch in the Bay Area, but if you don’t want to do something then just don’t apply for it. There are plenty of backend jobs available. The best way to get afoot in the door is through referrals and if you answer the standard CS interview questions you should be able to get a job pretty easily.
|
# ? Nov 1, 2017 20:47 |
|
Obviously I'm not much help Alpha but I want to wish you good luck. How did you find the 1 year cert program? Georgia IT has a Machine Learning Specialization in their online masters program (total cost is like 7k or something insanely low). I have no idea really, but I'd think if you want to do machine learning you'd be more competitive with a masters in it AND a few years of work experience in any kind of development. That's really just a guess, though.
|
# ? Nov 1, 2017 21:21 |
|
Thanks everyone. I guess I'll drop machine learning for now and just drill CS and cracking the coding interview. That's pretty much all I should be doing, right? I've been doing a few problems on Hackerrank in JAVA per day recently, but I wanted to know if it was worth switching to python and just whiteboard with that. I'm thinking its much easier to whiteboard and prototype with, with a lot less language memorization. I sometimes freeze up momentarily under pressure, like in exams, and forget JAVA's dumb arbitrary stuff. I just don't have too much practice with python, and I don't know how to solve complex loop problems with it yet. Hellblazer187 posted:Obviously I'm not much help Alpha but I want to wish you good luck. How did you find the 1 year cert program? Thanks, I appreciate it! The cert was pretty cool--it's graduate level and meant to be nearly equivalent to a bachelor's for people with no experience, so it's like Boot Camp Plus. It was meant to feed directly into a Master's in CS program, but due to personal issues I couldn't continue that path right now (thus my current dilemma). I'll check out Georgia IT. I know if I want to pursue a career in AI/DL that I have my work cut out for me and have to take on a ton of extra stuff.
|
# ? Nov 1, 2017 23:08 |
|
AlphaKeny1 posted:Thanks, I appreciate it! That sounds pretty great. Was it local to you or an online thing? I'm astounded they can offer a MS in computer science for $7k when the next cheapest is like $30k. No idea of the quality of the education! http://www.omscs.gatech.edu/
|
# ? Nov 1, 2017 23:53 |
|
I'll be honest, as a hiring manager who hires for machine learning I have very little faith in most machine learning or data science programs. If you want to be have the academic credentials for it you get a PhD in stats. If you don't want to do that, do kaggle competitions until you're like in the top 25%. (Don't be the guy who put that he was in the top 98%) I know jack poo poo about machine learning and the people who come out of those programs know even less.
|
# ? Nov 2, 2017 00:00 |
|
AlphaKeny1 posted:JAVA fyi java is not an acronym.
|
# ? Nov 2, 2017 00:44 |
|
ohgodwhat posted:I'll be honest, as a hiring manager who hires for machine learning I have very little faith in most machine learning or data science programs. If you want to be have the academic credentials for it you get a PhD in stats. If you don't want to do that, do kaggle competitions until you're like in the top 25%. (Don't be the guy who put that he was in the top 98%) I mean, that depends on the application? There are a lot of machine learning-adjacent positions like planning domains where, while nice, being really good at tuning classification accuracy a la Kaggle isn't really as important. Stats is stronger for, like, business analytics stuff, but that's not the only domain ML is used in.
|
# ? Nov 2, 2017 00:46 |
|
FamDav posted:fyi java is not an acronym. It doesn't stand for "JAVA Ain't Very Acronymic"?
|
# ? Nov 2, 2017 00:51 |
|
CPColin posted:It doesn't stand for "JAVA Ain't Very Acronymic"? I was trying to figure out how to make a recursive acronym joke and I couldn't come up with one.
|
# ? Nov 2, 2017 00:52 |
|
AlphaKeny1 posted:Thanks everyone. I guess I'll drop machine learning for now and just drill CS and cracking the coding interview. That's pretty much all I should be doing, right? You could do Python but if you don't know it why not stick with Java? Its a bit simpler but honestly noooot that different. Learning python's differences will take you longer than correcting your java mistakes. And depending upon the place you interview, I know I just interviewed somebody last week who would rather white-board in python and I let her, but then a co-worker asked her to try to work in Java if she could. When whiteboarding things aren't going to be perfect, minor syntax errors are forgivable. We wanna see you understand the algorithms and design, not the syntax necessarily. The IDE will do that poo poo for you for the most part. But if you're really struggling with Java syntax, you shouldn't necessarily just blow that off... odds are that means you're kinda shaky on some OOP concepts or such, and that'll come out in interviews. If you're shaky on java, work on java more. Write a quick program to get yourself more confident. I mean, you could do the same for python, write a quick program to learn it, but... yeah. But yeah Cracking the Coding Interview is the poo poo. If you can confidently answer anything in that book you should be golden. And yeah, I would agree with that and would have said the same; focus on getting your foot in the door with basic code monkey programming. Once you're getting real world experience on your resume on a team, then you're good, then go do some machine learning on your own time, and then next year or whatever you can see about hopping to a different team at the same company that does machine learning, or leaving and getting a new job that does. Zaphod42 fucked around with this message at 01:49 on Nov 2, 2017 |
# ? Nov 2, 2017 01:46 |
|
Linear Zoetrope posted:I mean, that depends on the application? There are a lot of machine learning-adjacent positions like planning domains where, while nice, being really good at tuning classification accuracy a la Kaggle isn't really as important. Stats is stronger for, like, business analytics stuff, but that's not the only domain ML is used in. I'm not sure where you're coming from with this. For one, there are a number of kaggle competitions and they can be amenable to a wide swath of ML techniques; writing them off as just "tuning classification accuracy" seems absurdly reductive. Two, I'm talking about what will get you hired, and having demonstrable experience applying machine learning techniques is valuable and kaggle provides a good way to do that. Three, I'm clearly not saying stats is required for ML, I just think it's a good basis. However, I will say that of the hundreds of applicants I have dealt with, those with degrees in ML (not other degrees with specializations in ML) were frequently little more capable than those without any ML background at all. What I'm saying is that those programs are not particularly useful as far as I've seen. Maybe it impresses other employers though?
|
# ? Nov 2, 2017 03:10 |
|
I don't doubt that someone will hire you if you go through a machine learning bootcamp, but it probably won't be me.
|
# ? Nov 2, 2017 03:14 |
|
Speaking of which, what's the next big buzzword and can we start a boot camp for it?
|
# ? Nov 2, 2017 03:27 |
|
Blockchain.
|
# ? Nov 2, 2017 03:34 |
|
Moat.
|
# ? Nov 2, 2017 03:37 |
|
Proficient at utilizing blockchains to swing across moats.
|
# ? Nov 2, 2017 03:40 |
|
ultrafilter posted:Blockchain. Which reminds me: Wasn't IBM writing some banking ledger software that would take advantage of the blockchain to have immutable and traceable transactions? Anything come out of it? I always thought that this would be a great way to use the technology, not the lovely bitcoin mining. Storing historical data seems perfect for it.
|
# ? Nov 2, 2017 03:41 |
|
I legitimately could see a blockchain bootcamp existing. It would sell blockchain as the future, but would only accept tuition in cash.
|
# ? Nov 2, 2017 03:43 |
|
Volguus posted:Which reminds me: Wasn't IBM writing some banking ledger software that would take advantage of the blockchain to have immutable and traceable transactions? Anything come out of it? I always thought that this would be a great way to use the technology, not the lovely bitcoin mining. Storing historical data seems perfect for it. Yeah, i would like to know this too. It seems like the real untapped (to my ignorant brain) potential of the blockchain. edit: ^ https://www.transformationworx.com/ < there it is.
|
# ? Nov 2, 2017 03:44 |
|
Goldman Sachs is talking very seriously about blockchain.
|
# ? Nov 2, 2017 03:51 |
|
Zaphod42 posted:And yeah, I would agree with that and would have said the same; focus on getting your foot in the door with basic code monkey programming. Is web dev the easiest place to break in with basic code monkey stuff?
|
# ? Nov 2, 2017 04:10 |
|
I regularly get notifications on blockchain and cryptocurrency poo poo, especially on hot new online courses. I bet if you write a Medium post about 'Your Path to Blockchain Career Fast', you'll get so much applause and get linked on Forbes.ohgodwhat posted:I'll be honest, as a hiring manager who hires for machine learning I have very little faith in most machine learning or data science programs. If you want to be have the academic credentials for it you get a PhD in stats. If you don't want to do that, do kaggle competitions until you're like in the top 25%. (Don't be the guy who put that he was in the top 98%) I was literally going to try the Titanic dataset today to get started but this thread has convinced me otherwise. But that's where I'll start when I pick it up again down the line. Zaphod42 posted:But yeah Cracking the Coding Interview is the poo poo. If you can confidently answer anything in that book you should be golden. Thanks, you're right. I'll stick to java. When I hand write a solution in java and type it up to see if it compiles there is always tons of syntax errors. I'll just have to suck it up and practice harder. Hellblazer187 posted:That sounds pretty great. Was it local to you or an online thing? FamDav posted:fyi java is not an acronym. I don't know why I wrote it that way.
|
# ? Nov 2, 2017 04:13 |
|
ohgodwhat posted:I know jack poo poo about machine learning and the people who come out of those programs know even less. Star War Sex Parrot fucked around with this message at 04:26 on Nov 2, 2017 |
# ? Nov 2, 2017 04:23 |
|
MSML is a master's program, yeah?
|
# ? Nov 2, 2017 04:26 |
|
ultrafilter posted:MSML is a master's program, yeah? Star War Sex Parrot fucked around with this message at 04:42 on Nov 2, 2017 |
# ? Nov 2, 2017 04:29 |
|
It's certainly possible our recruiters aren't reaching the good ML programs.
|
# ? Nov 2, 2017 04:35 |
|
The master's programs in data science that I've seen are pretty uniformly blatant cash grabs. A program in machine learning might not be, but if it's not really heavy on statistics, I'd be suspicious.
|
# ? Nov 2, 2017 04:43 |
|
Tezzeract posted:And no one really knows how Deep Learning works other than to throw GPU processing power and more layers at problems I forgot to post this
|
# ? Nov 2, 2017 05:03 |
|
|
# ? Jun 5, 2024 04:44 |
|
lifg posted:I legitimately could see a blockchain bootcamp existing. It would sell blockchain as the future, but would only accept tuition in cash. Like the old saying goes: sell shovels
|
# ? Nov 2, 2017 14:51 |