While I was studying at University, I’ve decided that I’m going to become
Anincreasing number of Twitter and LinkedIn influencers preach why you should start learning Machine Learning and how easy it is once you get started.
While it’s always great to hear some encouraging words, I like to look at things from another perspective. I don’t want to sound pessimistic and discourage no one, I’ll just give my opinion.
While looking at what these Machine Learning experts (or should I call them influencers?) post, I ask myself, why do some many people wish to learn Machine Learning in the first place?
Maybe the main reason comes from not knowing what do Machine Learning engineers actually do. Most of us don’t work on Artificial General Intelligence or Self-driving cars.
It certainly isn’t easy to master Machine Learning as influencers preach. Being “A Jack of all trades and master of none” also doesn’t help in this economy.
Why do so many wish to learn Machine Learning?
While I was studying at University, I’ve decided that I’m going to become a Machine Learning Engineer. It seemed hard, challenging and most importantly fun. Before that my wish was to become an iOS game developer.
If someone would show me a workday from an ML engineer, maybe I would stick with iOS game development. Don’t get me wrong, I’m really happy with my career, but the career choice wouldn’t be as black and white as it was.
Why is that? Because you can get the same amount of fun writing an iOS game as with training a Machine Learning model… or developing a backend application… or a frontend application. All of the above can become challenging (just ask engineers at top Tech companies).
While in University, my thinking was:
Machine Learning seems hard, so it’s going to be easier to get a job. I’ll get higher wage. It is more future proof (web development will soon be automated) and it is fun.
My thinking was wrong. So allow me to explain each of the statements above.
1. Machine Learning seems hard
Most internet influencers preach: Starting with Machine Learning is really easy. You just download the Titanic dataset, copy 10 lines of Python code from a tutorial and you’ve started with Machine Learning.
While that’s true, it’s hard to imagine that someone would pay you for that knowledge. So you need to go levels deeper.
And levels deeper is where it gets hard. Having a great mentor is really important so that you don’t need to figure out everything on your own. Getting a good internship is also a great way to grow as an engineer.
I wish someone would tell me that at the beginning of my career. I had to put in considerable hours to keep up with my peers who worked in other areas of Computer Science.
Why? Well, It’s easier to get a mentor for frontend (backend or mobile) development because there’re more people doing it.
2. Easier to get a Machine Learning job
One thing is for sure and I learned it the hard way. It is harder to find a job as a Machine Learning Engineer than as a Frontend (Backend or Mobile) Engineer.
Smaller startups usually don’t have the resources to afford an ML Engineer. They also don’t have the data yet, because they are just starting. Do you know what they need? Frontend, Backend and Mobile Engineers to get their business up and running.
Then you are stuck with bigger corporate companies. Not that’s something wrong with that, but in some countries, there aren’t many big companies.
3. Higher wages
Senior Machine Learning engineers don’t earn more than other Senior engineers (at least not in Slovenia).
There are some Machine Learning superstars in the US, but they were in the right place at the right time — with their mindset. I’m sure there are Software Engineers in the US who have even higher wages.
4. Machine Learning is future proof
While Machine Learning is here to stay, I can say the same for frontend, backend and mobile development.
If you work as a frontend developer and you’re satisfied with your work, just stick with it. If you need to make a website with a Machine Learning model, partner with someone that already has the knowledge.
5. Machine Learning is Fun
While Machine Learning is fun. It’s not always fun.
Many think they’ll be working on Artificial General Intelligence or Self-driving cars. But more likely they will be composing the training sets and working on infrastructure.
Many think that they will play with fancy Deep Learning models, tune Neural Network architectures and hyperparameters. Don’t get me wrong, some do, but not many.
The truth is that ML engineers spend most of the time working on “how to properly extract the training set that will resemble real-world problem distribution”. Once you have that, you can in most cases train a classical Machine Learning model and it will work well enough.
I know this is a controversial article, but as I already stated at the beginning, I don’t mean to discourage anyone.
If you feel Machine Learning is for you, just go for it. You have my full support. Let me know if you need some advice on where to get started.
But Machine Learning is not for everyone and everyone doesn’t need to know it. If you are a successful Software Engineer and you’re enjoying your work, just stick with it. Some basic Machine Learning tutorials won’t help you progress in your career.
The aim of this article was to give a critical view that you usually don’t hear from influencers.
Before you go
Here are a few links that might interest you:
Some of the links above are affiliate links and if you go through them to make a purchase I’ll earn a commission. Keep in mind that I link courses because of their quality and not because of the commission I receive from your purchases.