Read on Twitter
Transitioning from Software Engineering to Machine Learning.
I'll tell you my story.
It might work for you. It might not.
Hopefully, it gives you another perspective. Hopefully, it helps.
(2 / 14)
It might work for you. It might not.
Hopefully, it gives you another perspective. Hopefully, it helps.
(2 / 14)
Many people see "Software Engineering" and "Machine Learning Engineering" as two completely different specialization areas.
There are many differences, for sure.
But I personally like to think about them as a single, fluid, all-encompassing position.
(3 / 14)
There are many differences, for sure.
But I personally like to think about them as a single, fluid, all-encompassing position.
(3 / 14)
Regardless of title, the ultimate measure of success is your ability to produce good software.
Your skills will always be centered around something specific, but in the end, the only thing that matters is working software.
(4 / 14)
Your skills will always be centered around something specific, but in the end, the only thing that matters is working software.
(4 / 14)
I like to look at "building software" as a spectrum instead of many individual silos.
Your responsibilities at Company A might look very different than your responsibilities at Company B (under the same title.)
(5 / 14)
Your responsibilities at Company A might look very different than your responsibilities at Company B (under the same title.)
(5 / 14)
A Web Designer at Company A may be responsible for providing designs to a Front-End Developer.
At Company B, the same Web Designer may be responsible for the designs, HTML, and CSS, while the Front-End Developer may focus on the JavaScript code.
(6 / 14)
At Company B, the same Web Designer may be responsible for the designs, HTML, and CSS, while the Front-End Developer may focus on the JavaScript code.
(6 / 14)
Many Machine Learning professionals don't come from a Software Engineering background.
They are laser-focused on creating models. Their work starts with data and ends with a good performant model.
Someone else cares about productizing those models.
(7 / 14)
They are laser-focused on creating models. Their work starts with data and ends with a good performant model.
Someone else cares about productizing those models.
(7 / 14)
This has helped enforce the notion of "silos."
"You are a Software Engineer. Therefore you do A, and B, and C."
"And you are a Machine Learning Engineer, so you focus on D and E."
(8 / 14)
"You are a Software Engineer. Therefore you do A, and B, and C." "And you are a Machine Learning Engineer, so you focus on D and E."(8 / 14)
Personally, I had years building software before starting with Machine Learning.
As I started down that path, I incorporated more and more of the new skills in the different projects I was involved in.
Slowly. Step by step.
(9 / 14)
As I started down that path, I incorporated more and more of the new skills in the different projects I was involved in.
Slowly. Step by step.
(9 / 14)
We were building a dashboard to show some metrics to executives of a company.
I was able to forecast future trends based on past data.
That was one of my first "crossover" assignments: software development work + machine learning.
(10 / 14)
I was able to forecast future trends based on past data.
That was one of my first "crossover" assignments: software development work + machine learning.
(10 / 14)
I didn't get promoted. I didn't switch roles.
I was the same developer, but now with an extra set of skills.
Step by step, I started applying what I was learning to my current job.
(11 / 14)
I was the same developer, but now with an extra set of skills.
Step by step, I started applying what I was learning to my current job.
(11 / 14)
Over time, I started tackling harder problems. The shape of the work that hit my plate morphed.
Nothing happened overnight. This "transition" has been taking place for years.
(But it's not a "transition," really. It's more of an "expansion.")
(12 / 14)
Nothing happened overnight. This "transition" has been taking place for years.
(But it's not a "transition," really. It's more of an "expansion.")
(12 / 14)
I understand that I've been fortunate: not everyone has the chance to "expand" their workload.
Unfortunately, many people need to switch roles (even companies) to focus on machine learning problems.
(13 / 14)
Unfortunately, many people need to switch roles (even companies) to focus on machine learning problems.
(13 / 14)
So there you have it.
I'm usually hesitant to call myself a "Software Engineer" or a "Machine Learning Engineer."
I like to see my role from a different perspective.
I build software. That's what I'm good at.
(14 / 14)
I'm usually hesitant to call myself a "Software Engineer" or a "Machine Learning Engineer."
I like to see my role from a different perspective.
I build software. That's what I'm good at.
(14 / 14)
