Math PhD with 1.5 years experience, not sure whats a good fitFeb 11
When I was finishing my phd I had a very poor understanding of what I was worth and took a job without much searching or interviewing. I took a research scientist position at a small firm and have seen two deep learning projects to production, so I figured it was a good time to look around and see what I might do if I was elsewhere.
I am not really sure I know what kind of role I would be best suited for.
I am not sure where I would fit or how to prepare for the interviews. Should I grind leetcode or make sure I can prove hoeffding? Most preparation sites seemed aimed at SWE positions but is there something that is more up my alley?
Second question - what is the difference between Research Scientist / Applied Scientist / Machine Learning Researcher?
Current TC is 150
- Intel KTPu74These are just names. Work that you do is agnostic of what the job title says. In my group everyone suddenly got designated as data scientist, while some people don’t even know what linear regression is .
- I disagree with the first post. It might be like that Intel but not at some other top companies. Lets start with different titles.
1. Applied scientist, ML engineer: synonymous with each other as far as I know. Some knowledge of ML, strong coding (not as strong as SWE but still good enough to clear 3 coding rounds at Google). Good if you want to implement an ML based product and not care about publishing much. Also, you may not get to innovate.
2. Data scientist: Be cautious of this title. In some companies, it means ML engineer and in some companies it means glorified data analyst.
3. Research scientist in a non-research lab: sometimes it is genuinely a research scientist position means you don’t necessarily work on something that will make it to a product in a year or so. You innovate and the focus is one publishing. Since you are not working on production code, your success is measures by patents and papers. A high pressure position. I have worked in a research lab before. Sometimes, it might also be a glorified applied scientist position.
4. Research scientist in a research lab: definitely a genuine research scientist
You decide what you want to do.
- You can try Amazon. Applied Scientists need to pass both sde-1 level coding bar (LeetCode) + ML bar (for this I’d recommend going through Amazon’s free ML course).
For research scientist, just ML knowledge and some coding skill would be enough!
- Im not sure anyone can tell you what role you're best suited for. What is it that you'd like to spend your time doing? Applied research or straight coding? My guess is more the former which would probably have some coding involved still unless it was MSR or similar.