Is an MFE worth it if I'm a Senior Engineer at FAANG?

Hi all,
I'm currently a Senior Research Engineer at a FAANG company (IC5 on track for IC6) pulling in close to $450k+ working in DL/ML/AI research. I used to be finance, and even got admitted to Baruch's MFE back in 2014 before I made the switch to mega-cap tech to work on self-driving and more recently on VR tech. Now I'm looking to move back to finance, and thinking about going back for an MFE because: mostly I like trading, and I like math.
The real question is: Is it worth switching at this stage in my career? How much value would a trading firm put on the engineering/ML experience combined with the MFE. I was planning on a part-time MFE (while working) till the last semester, and then taking a few months off before switching to a trading firm but I'm worried that I'd be walking away from a ton of lost earnings and would end up starting back from square 1 (>50% comp cut) in terms of earnings.

Is there an equivalent to: levels.fyi for quants?
 
if you're interested in research based roles, i'd suggest start applying to firms. the hedge fund i work at regularly interviews people like you, and i'm pretty sure there are many others like that. i've seen people not being asked any finance questions as long as their background and skillset in ml/math/stats/cs is top notch.
 
No. Get a role in trading infrastructure/quant development first if you absolutely have to, do not take a step down lol. There's the perception that SWE -> quant is a hard ask, but I don't think this applies to SWEs working on ML research stuff as you are. Only bad shops emphasize perception and job title as opposed to what you actually do day-to-day.

Quant comp is converging towards levels.fyi anyway (maybe with heavier tails for the absolute whiz outliers) so just use that, it'll suffice
 
Last edited:
I don’t know, I moved to faang in AI/ML as a contractor because I had trouble getting interviews out of Baruch. It’s been 8 years since I graduated with 0 quant interviews. Ok maybe I lie, I had 2 interviews at max. I would say a mfe isn’t worth it. I’m almost pushing 40 now. Sigh freakin graduated in 2014 when I was 29. And 0 career development.
 
I don’t know, I moved to faang in AI/ML as a contractor because I had trouble getting interviews out of Baruch. It’s been 8 years since I graduated with 0 quant interviews. Ok maybe I lie, I had 2 interviews at max. I would say a mfe isn’t worth it. I’m almost pushing 40 now. Sigh freakin graduated in 2014 when I was 29. And 0 career development.
Baruch is the top of the funnel as far as MFE programs go and 2014 was not a bad time to be trying to be getting in (try after 2017-2018...), and if that still sucks then that means the MFE skill set is commodified and you will need to have something to show other than it to get good offers (certainly this is true for the MFE students that do get these offers, e.g. have great software engineer experience before the MFE, do well in kaggle competitions, etc)
 
When I was working at Morgan Stanley as a contractor, I was working with few strategists. Honestly I felt they lacked any programming skills and they barely had any ml knowledge. Writing close to 1k lines of code in a ipython notebook to send it do me to productionalize. I was like WTH? This clown is a strategist? SMH. Then again Morgan Stanley’s wealth management tech team is a abject joke. They are atleast 20 years behind fb in terms of their tech stack. Fb was the best company I worked for hands down.
 
When I was working at Morgan Stanley as a contractor, I was working with few strategists. Honestly I felt they lacked any programming skills and they barely had any ml knowledge. Writing close to 1k lines of code in a ipython notebook to send it do me to productionalize. I was like WTH? This clown is a strategist? SMH. Then again Morgan Stanley’s wealth management tech team is a abject joke. They are atleast 20 years behind fb in terms of their tech stack. Fb was the best company I worked for hands down.
This is extremely common in the field, unfortunately. It's not only annoying, but creates significant hidden, operational risk for companies because the person tasked to productionize the code mess is not the same person who did the research, and disconnects often occur between the two.

The QN Python/Data Science course (Python for Finance with Intro to Data Science, disclaimer: I'm the originator) instills coding best practices in Python prior to introducing IPython (JupyterLab) for data/ml research, with the hope that even non-developers leave the course with a proper approach to coding even when doing adhoc data research.
 
Last edited:
When I was working at Morgan Stanley as a contractor, I was working with few strategists. Honestly I felt they lacked any programming skills and they barely had any ml knowledge. Writing close to 1k lines of code in a ipython notebook to send it do me to productionalize. I was like WTH? This clown is a strategist? SMH. Then again Morgan Stanley’s wealth management tech team is a abject joke. They are atleast 20 years behind fb in terms of their tech stack. Fb was the best company I worked for hands dow
I think the guy writing 1k lines of shi**y code is not paid to write production level code. Writing Production level code is a different skillset altogether compared to programming python notebooks.
 
Top