Need internship referral in ML
I’m a rising sophomore interested in interning during the 2019 summer in the bay/seattle. I know it’s early but I was hoping somebody would refer me. I spend a lot of time playing with Keras and TF. I’ve made several different deep learning image classifiers (more than MNIST) and assist research in informatics (visual adversarial attacks on NN). I was hoping something for Microsoft/goog/nvidia/Apple/Amazon etc. TC: 4K/ month + free oatmeal for breakfast at cafeteria (with the brown sugar) Lmk if I’m in way over my head or something
Coursework will matter more since you're a sophomore. What's the largest dataset you've worked with? Any experience with FlumeTFX, production scale ML?
Never worked with TFX or any production scale ML. I recently realized working with Production scale stuff would be critical. Relatively small datasets (limited comp. power). I intend on doing more kaggle challenges on AWS with the intent of making my submissions work like pipelines that aren’t restricted to data set size.
As a tip, look for research projects and classes where the goal is large scale data analysis so you get projects in that domain. Maggie challenges are great if you place top 3 but otherwise you're better off winning your local university data science cups (much easier to win less competition and they look even better than Kaggle).
Focus on data structures. Being a sophomore, it’s highly unlikely that you will get to develop an ML model right away. They might put you into data labeling (not easy as it seems should it be done correctly) or in the best case implementation. Make a GitHub repo, put all your work there. If you want to play with big datasets, for image domain, ImageNet is the way to go, but not necessary imho. I would find it crazy if interviewers expect you to have worked on production scale projects. Make a webiste if you are into it or at the very least, make your LinkedIn impeccable. Contact recruiters and managers. Classification and GANs are trending right now so that should help. Take a good paper in GAN domain and implement it, if you achieve it.. that is something to brag about! I wish you good luck! You are very proactive given you are so young. P.S. I really liked your TC and free oatmeal part. Made me smile.
Thank you for the feedback. I’ve implemented several primitive versions of research papers but I’ll see if I can take that a step or two further. The funny thing is I eat oatmeal at home every day too!
How you project yourself is very important now-a-days. So pay attention to how you explain your solution. I have seen some people fail because they couldn’t explain their solution.
Ah keen youngins ahead of the game already. Keep working hard and get that PhD
Oh stop it you
Being a sophomore, just machine learning knowledge and the familiarity of relevant frameworks is not going to be enough to crack the companies you have mentioned. You need to show good skills in Data Structures and Algorithms as well as most of the interviews will focus on that primarily. You can always express your interest in the specific fields later on. Leetcode / GeeksforGeeks should help. Though your side projects will add value as well :)
I’ll hop on the leetcode bandwagon then. I’ve done quite well in my course work but I’ll need work if i have to answer any serious Q’s on the spot