2

Validating career pivots (SWE to MLE) with an online masters program?

Profile picture
Senior Software Engineer [E5] at Meta7 months ago

Hi all, I'm currently working on GenAI server infra at Meta as an IC5. I'm interested in the space and have had a long standing interest to dig deeper into ML. At a high level, I'm considering the following:

  1. Positioned well for E6 scope and exploring interest in people management (M1)
  2. Taking an online masters program to brush up on math, ML foundation skills. This can better validate my interests. Alternatively, I can lean into online resources like Cousera, deeplearning.ai
  3. Recruit internally as an E5 to a relevant team. Generally internal mobility empowers people to explore and learn new things. However, it helps to have relevant experience.

For online masters programs, I've heard of great things from Georgia Tech, with a ~$7k online program. Alternative options include UT and Stanford, and the latter is ~$60k. Rahul made a great video about his experience at Stanford masters, but this would rather be a part time commitment, and I already have many YOE at Meta for my resume.

I'd also love to hear your general thoughts on navigating career moves into an area with existing, highly talented pool.

82
2

Discussion

(2 comments)
  • 2
    Profile picture
    Founding ML Engineer @ Lancey (YC S22)
    7 months ago

    Hi OP, when you say dig deeper into ML, what exactly do you mean? are you looking for more R&D role? It seems like you're currently working on platform.

    I'm sure at Meta the core ML is reserved for PhDs. But I've heard of MLEs who do more engineering related work of ML can be done with SWE experience. Open AI themselves said that they value SWE expertise more than ML expertise because its easier to teach someone to do ML than write good code.

    Generally I think it's better to try and transition and learn on the fly rather than get another degree especially for ML. ML roles are either core ML which at big tech you'll need PhDs because they're specialized roles or they are supporting ML which is basically a different flavor of SWE.

    The best option imo is to first try to break into ML by applying or trying an internal transition. Soon you'll realize what you're missing for your career transition. Best case scenario you dont need a masters for what you want.

  • 1
    Profile picture
    Tech Lead @ Robinhood, Meta, Course Hero
    7 months ago

    I'm personally not a fan of doing a Master's unless it's a hard requirement as we talk about here: "Is an online Master's Degree in Computer Science worth it for a non-CS Bachelor's degree holder with 2 YOE?"

    My wife did Georgia Tech OMSCS with an ML focus, and it hasn't led to any career advancement for her. It took her a ton of time, a lot of the material was outdated (academia is almost always behind industry), and she doesn't remember most of what she learned there.

    In your situation, a Master's makes even less sense IMHO as you are in a better position than 99% of engineers on Earth by being a high-performing E5 at Meta. It's a much better use of your time to:

    1. Push towards E6 - Follow the advice here: [Course] Nail Your Promotion As A Software Engineer
    2. Network with AI/ML engineers - You literally work at one of the world's leading AI companies, so this should be relatively easy. Get to know the MLEs and Research Scientists within your organization. Have ☕ chats and try to understand their day-to-day. Use the tactics from our networking masterclass to get them to like you: [Masterclass] How To Build Deep Relationships Quickly In Tech
    3. Work on deeper ML projects - You are already working on GenAI server infra. As an E5 (and especially as an E6), the expectation is to be able to find your own scope. Can you expand the scope of your current projects to take on tasks that are deeper within the AI/ML space instead of being more within the overall back-end space? If you've done #2 well, that will help a lot. You can ask those engineers if they have any scope they can offload to you (this is a win-win as they get to delegate work that's straightforward for them but a great learning experience for you).