Full Stack to AI/ML engineer
Can someone help me with a roadmap to switch to AI/ML engineer . As in what should I learn and upskill upon .. to get a role at some of the top companies in this domain
Thanks In Advance
Discussion
(5 comments)- 1
3 months agoFounding ML Engineer @ Lancey (YC S22)Hey OP, see my answer on this thread here: https://www.jointaro.com/question/48WePYRfWEXgc5itTYUY/new-grad-backend-greater-mle/
The main insight here is find SWE tasks on ML teams. I really dont suggest getting a masters or taking specialized training for this. It's far easier to switch internally than to make a double change (change companies + role in 1 go) in this market
Happy to chat about it on slack if you have more questions.
- 1
3 months agoTech Lead @ Robinhood, Meta, Course HeroThe main insight here is find SWE tasks on ML teams.
This is great insight. Organically expanding your scope this way will be the lowest-friction option.
- 13 months agoFounding ML Engineer @ Lancey (YC S22)
Adding to this, ML has a tonnnn of breadth. the modeling aspect (which is what gets all the hype) is like 5% of the employees on an ML team and 95% is everything needed to get a model to production and im not even talking about SWE teams using the output of ML models in their work.
A lot of times the modeling aspect is gatekept to people with PhDs because at the scale of big tech where things are already hyper optimized you need a team of PhDs to make even the smallest improvement to the models
- 04 months agoTech Lead/Manager at Meta, Pinterest, Kosei
IMO it's hard to give a roadmap since most lateral career moves (e.g. going from one type of engineer to another) is so dependent on your current situation like company, skill set, and who you know.
I would recommend leveraging your existing skill set (there's a ton of overlap between a good AI engineer and a good generic software engineer), and ideally find work within your company that moves you toward AI.
The other option is to do a "reset" career move, e.g. do a masters or PhD, but I wouldn't recommend that in general.
Some really great answers about the day/work of an AI engineer here: Moving to AI/ML from web development?
- 03 months agoTech Lead @ Robinhood, Meta, Course Hero
Here's a high-level roadmap for any tech stack pivot as an engineer (the principles are generic): "How to transition from back-end development to distributed systems?"
tldr; Switching roles internally is by far the best way, and the main way I've seen success stories.
If you're doing outside learning, it's incredibly hard (but definitely possible, especially with Taro) to break into AI/ML as millions of other engineers have the same idea. Your options are:
- Make a crazy successful side project - I'm talking 10,000+ users minimum or 1,000+ GitHub stars.
- Become an AI/ML open-source master - This means contributing dozens of high-impact PRs minimum to core AI/ML libraries (5,000+ GitHub stars).
- Get a PhD from a top school - I'm talking Stanford level. A Master's won't cut it for this field in this market as it's too accessible. Most Master's students I know right now are struggling and have effectively 0 chance to get into a place like OpenAI.
Here's another good thread about the topic: "Is it worth transitioning to become a Machine Learning Engineer?"
