You can do that from home?”Before, I had a succinct, time-tested definition that relied on explaining my role in terms of who I worked with, who I sat near in the office, what software I used.
But now I had to explain what I did in terms of the value I added.
There’s no Venn diagram that explains me sitting at home on my laptop.
This process of trying to explain myself to all of these people really helped me understand what it is data scientists do for the organisations they work in.
And this understanding, in turn, helped me communicate the value of my skills to clients.
What is data science?It’s not about languages or frameworks or algorithms.
It’s not about org-charts and it doesn’t matter who your line manager is.
It isn’t a subfield of AI or a superset of machine learning.
Data science is making predictions to help people or machines make decisions.
This definition covers a lot of ground.
It rules out data analysis and business intelligence work by including prediction-making as a prerequisite.
And, more importantly, it gives data science a point — If you’re not helping people or machines make decisions, you’re not doing the work of a data scientist.
It’s important to step back from the technical, tool-filled, jargon-filled answer because doing so makes it easier to see where data science jobs come from.
Where data science jobs come fromCompanies don’t hire data scientists because they like having them around.
They hire us because we can help them make decisions.
But even that isn’t enough impetus for the average company to pay a data scientist’s salary.
Data science jobs get made when a particular decision needs to be made or a decision-making process needs to be improved.
“We need a recommendation engine” is really “we need help deciding what to recommend.
”“We need to predict outliers” is really “we need to decide who to investigate for fraud.
”“We need a cutting-edge computer vision algorithm for object detection” is really “we need a computer to make these decisions because having our staff make them hundreds of times a day is destroying their morale.
”These forces are the ones that actually create data science jobs.
How to get a data science jobBut how does this affect you, the person trying to get a data science job?It affects you because it means that when you rush to prove how qualified you are, how well-studied, how technically capable, that you may be missing the point.
It’s easy to do.
You get excited when you get the call, when they invite you for an interview.
The advertisement says they want to use the latest and greatest from your specialty.
You’re already picturing what it would look like in your portfolio.
And so you go, and you try to best to convince them you can do what they asked.
You can build that recommendation engine, you can deliver the cutting-edge computer vision algorithm.
And then you don’t get called back.
So what the hell happened?You convinced them you know the solution — there’s no doubting that.
But you didn’t convince them you know the problem.
Data science problems don’t occur in a vacuum, they’re a part of a business, a part of a process, a part of people’s lives.
When you go to an interview, the person sitting across from you is waiting to hear something in particular.
They’re waiting for you to ask what decision needs to be made.
So next time you have a data science interview, ask why.
Why do you need a cutting-edge computer vision algorithm?Why do you need me?Why do you need to do this now?If you ask these three questions with confidence something else will happen.
You’ll occupy a different place in your interviewer’s mind.
You’ll be the problem solver, the expert, the person who understands the problem.
And you’ll stop being the commodity technician.
That’s a transformation that’s worth three questions.
Isn’t it?Looking to take the next step in your data science career?.Take the quiz.
Originally published at carldawson.
net on March 26, 2019.
.. More details