Does it make sense for me to choose a specialization like ML? If so, what's the impact?

Question

Does specialization matter? I have a bunch of friends working in Big Tech, and some of them at my level started growing in 1 specific area. For example, one of them started going deep into machine learning. Do you need to make that decision when you’re a mid-level/senior engineer and what’s the criteria?

Adding on to this, I see a lot of noise in the market about various degrees and certifications to help you with specializations. Things like web3, ML, or other hot topics. This, combined with what I see my peers doing, makes me feel like this is a worthwhile or even the optimal thing to do.

As a counter-example, I have a friend who chose to specialize in ML and they’re job hopping a lot as they haven't been able to find good ML work at any of their recent companies.

Alex Chiou · Accepted Answer

Specialization really doesn't matter among the best tech companies as they just want good software engineers. The best software engineers have strong fundamentals and behaviors, which are generic and independent of tech stack. The companies that only want their SWEs to do very specific technical things tend to treat their SWEs as commodity coders and are weak overall. There are 2 paths here:   You enjoy the work you're currently doing: Then just keep doing it! It doesn't matter if you're working on an old and boring tech stack - You become a better SWE by having better depth. You can still add a ton of value being extremely good at a very traditional tech stack.   You don't really enjoy your current work and are soul searching: In this case, it can make sense to switch into one of these specializations. The overall point is that you should work on something you care about and enjoy. Don't take on a specialization simply because you want to work on the most cutting edge tech or to hyper-optimize your career opportunities; do it because it's a field you're legitimately passionate about. It is harder to fully switch into a specialization than many realize. For example, to really switch into ML, you can't just take a Udacity nano-degree and start applying to ML jobs. Even if you get a job, it's likely you'll get saddled with crappy ML work like data munging, which seems to be what happened to your friend. A lot of specializations require a PhD to really get compelling work within them, which takes a ton of time.  Related resources:  Is it worth transitioning to become a Machine Learning Engineer? I feel like senior engineers tend to focus on a specific area. How can I pick a specialization that's still relevant in 5-10 years? How to balance between specialization and pigeonholing? How To Find Your Path Among The Countless Paths In Tech

Does it make sense for me to choose a specialization like ML? If so, what's the impact?

Discussion