Fam, Engineer with 20 years of experience. Need some guidance on where to start with ML/AI. Did the obligatory Coursera AndrewNG certification. Besides theoretical knowledge not sure where/how to apply it. Any pointers ? Perhaps open source contributions or moon lighting elsewhere ? TC: 500K #engineering #software #swe
You have 20 yoe and don’t know how to learn something new? Don’t bother you won’t make the cut
:) Anyone can learn on the surface. Practical applications need battle hardening. Need to pick the right battles at the right time.
MS/PhD. Or try to become grandmaster in Kaggle through self-study.
Kaggle is a good place. Georgia techs OMSA is also good i heard... Maybe a lateral.move or picking up ML work at current work?
I can learn all the theory, trying to see if there is some way to get into the industry
What's your current line of work? It's easier to switch to the area by working first in the adjacent area rather than taking courses online
Full stack, distributed architectures
I’d study up on textbooks for theory, and start side projects for practice. Create a website that predicts things, or a Twitter bot that predicts stuff. Download public data sources and do fun analyses. Join a competition on Kaggle. The Coursera course is in Octave but please stick with python, sklearn, PyTorch. Try coding up a few ML algos from scratch: kmeans, logistic regression, a simple neural net. If you do all of that, you’ll be golden.
And a good Stats foindation helps too...
The stats part is right on the money, but incredibly hard higher math
Just skip ML, there is enough people in the field.
Well everyone is crowing so much about it, thinking it might be useful to know enough
If you’re serious about it: grad school. But why ML? It’s pretty saturated and you already make good tc.
fast.ai is a practical hands on first approach course https://course.fast.ai/ there's a book to go along with it https://course.fast.ai/Resources/book.html
Bootcamps