I'm a recent grad from a tier 2 university. I finished my masters in CS with 2 years of research in basic Natural Language Processing and thesis in Deep Learning. During my graduation, I interviewed with 33 companies and got to the final rounds (no on-sites, only video calls) of only 2 since I had very little time to prepare due to my thesis. I got dejected and accepted a basic Data Scientist job with 85k TC. I don't even do proper data science here as I mostly work on Tableau, data analysis and sometimes not even technical work. I've been working hard ever since, I finished deep learning specialization by Andrew NG, improved my basics in math, statistics and learning spark now. It might seem ambitious but I'm aiming for ML eng or data scientist position in top 10. I'd really appreciate any advice you can offer. Edit: People who work as a data scientist/ml engineer in the top companies - can you please explain what it really takes to get there?
You need more yoe, and you should have taken the hint after the first 5 or so.
But the fact that they knew my yoe and still interviewed didn't make me seem like giving up.
That just means you didn’t otherwise “wow” them.
Yoe?
1 yoe
Is there something particular you are looking for in an opportunity? Nature of work, domain, tech, or specific companies
What kind of questions did they ask during the on-site interviews?
I did not have typical on-sites, but I had a final round with a big fintech JPM on the phone. They've asked me stuff from algebra/calculus, stats, ml basics, coding, oops. I was just blown by the amount of ground I had to cover.
Seems like that's the problem then, you weren't prepared at all
Following
have you interviewed for data science stuff at the big companies that know what they're doing? If so, do you think they expect way too much in the interviewing process - especially compared to the SWE roles?
Thank you! Someone had to say it, there sure is so much variance to adapt to in the interviewing process. No matter how much I leetcode, I'd still get screwed coz the interviewer would ask me to code RNN from scratch in a coding round.
Leetcode
34 onsites in what amount of time? How much time do you spend preparing for each one? How much of that is research on the company and thinking through the ml problems at the company and how you’d approach them?
I got to only 3 final rounds. I spend about anywhere from 1 - 3 weeks max. I do research the company quite a lot and try to come up with interesting ideas to some of their problems, just to be prepared. Hardly get asked on those though.
For Applied Scientist/ Machine learning Engineers, it will be mixture of ml/ Nlp experience questions/ problems plus leetcode to check whether you can write good production code to model Processing in Systems
Thank you. I do realize that I need to be doing more leetcode than I do usually, based on the comments. But there are also occasional math/stats questions which do require quite some time to prepare for.
LeetCode a lot. It's the major indicator they look at with your YOE.
It's not evident from the resume