A Day in the Life of a Senior Data Scientist
Everything you never wanted to know about a day in the life of a senior data scientist. We take a voyeuristic look into the life of some guy called Nate.
Let me introduce you to Nate, or in a way, to myself! :)
Nate is a typical introverted but highly educated guy working as a data scientist at a big tech company. (Any similarity to actual persons is purely intentional!) He’s been a data scientist for six years and is considered somewhat of a senior data scientist. Let’s peek into one day of his life.
[09:45 AM]: Good Morning, Sunshine
His day starts bright and early at 9:45. You can’t say he’s sleep-deprived. His daily standup doesn’t begin until 11:00 a.m. because no one on his team can wake up that early. And the Bay Area traffic doesn’t die until 10:00 a.m.
Today is one of three days a week he has to commute to campus because ‘leadership’ thinks it’s necessary to collaborate in person.
[10:00 AM]: Today I’m Shuttling
Nate splashes some water on his face and hops on the shuttle to head to campus, where he catches more sleep on the shuttle. A nap a day keeps the doctor away.
[11:00 AM]: A Scrum Meeting
He gets to work on time for a scrum meeting.
What’s a scrum meeting? It’s a 5-minute meeting with your team to discuss three questions:
- What did you do yesterday? (Sleep.)
- What will you do today? (None of your business, you noisy prick!)
- Do you have any blockers? (Yes. You.)
It’s a pretty pointless meeting, but hey, a manager’s got to look like he’s managing!
[11:15] Meeting With a Junior Data Scientist
After the meeting ends, Nate jumps to the next one, which is a code review with a junior data scientist. So, who or what’s a junior data scientist?
Junior data scientists are those in the first few years of entry-level data science. Their job is to learn how to write production-level code and to produce outputs that are meaningful to the business.
Anything more than that is really just icing on the cake for these young, budding data scientists. Sounds easy. But it’s actually really tough.
This guy wasn’t a STEM major; in fact, he was a Psychology major turned data scientist. Imagine going to school to learn how to read books and write essays, then stepping into your career and being asked to write complex code and understand advanced statistics.
This is the stage where we weed out the wannabe data scientists. So we evaluate the juniors based on a few criteria:
- Ability to break down a problem and structure it in a meaningful way Can you take a vague, complex problem and break it into bite-sized pieces? Or do you need help doing that?
- 2. Ability to understand the business question and design approach that solves it. Most juniors just want to write overly complex code and use the newest, shiniest tools. That’s just asking for a lot of trouble. You’re over-engineering the problem and making it much more complicated than it needs to be. We’re not rocket scientists but scientists managing a social media platform.
- 3. Ability to write good, production-quality code Most people don’t know how to do this. You need to adhere to the best practices of the company, you need to consider scalability in your code, and you need to consider edge cases.
So Nate helps this junior clean up his code and offers some advice on making more meaningful progress on the project.
With that finished, Nate now has a one-on-one meeting with his manager.
[01:00 PM] One-on-one With a Manager
These one-on-ones are meant to be a discussion around new opportunities and career progression.
They can help propel Nate to the next stage of his data science career. These are valuable discussions when done right. But not today, oh, no! Today, there are a lot of issues and fires, and this one-on-one time will be used to discuss how to handle these issues.
It’s a useless 30-minute meeting, but Nate is glad to know that his manager is aligned with the way Nate will approach solving these fires. Being aligned and in sync with your manager and peers is a great soft skill Nate has developed throughout his career.
He understands that the next step in his career will emphasize soft skills like communication, alignment with cross-functional teams, and other non-data, non-coding things.
[02:00 PM]: Finally Doing Some Real Work
Now, it’s finally time to do some real work. Let’s start to code! Nate is starting on a new project where he will be building a new recommendation engine for a specific feature in a social media app that doesn’t really need any more social media manipulation.
He logs into his workspace, which uses tools made in-house, tools from AWS, his cloud service provider, and other third-party vendors.
He notices a problem: his permissions to access the data in his analytical environment are not working because it seems like the data engineers didn’t do their part. So, he puts in the request for the admins to take a look at it, and he can’t work until he has the data.
‘Luckily’ for him, there’s an extract of that data in some sort of sandbox environment. Nate takes a look and quickly realizes that:
- He needs to clean the hell out of this data.
- He doesn’t understand any of it.
He sets up a meeting with the data subject matter experts (SMEs) so he can understand the logic and the rules behind the data. In the meantime, he’ll clean the data the best he can.
This is the easy part, but the hard part is trying to understand the logic of the data so it can be used to build the machine learning model.
That said, the hours he set aside to code have become useless. He has to wait for help before he can do anything impactful. No modeling,
no advanced infrastructure design
or tooling
Nothing!
That’s typical. Much of Nate’s time is usually spent preparing rather than building. But he’s seasoned; he understands that prep work is 90% of the project, and building the model is the easy part.
[07:00 PM] Hasta mañana!
So, instead, he goes to grab some food from the cafeteria and answers emails until it’s time to go home.
On a typical day, he goes home at around 7:00 p.m. He feels a little unsatisfied because so many things were left undone today.
[08:00 PM]: Today is tomorrow. It happened.
He’s home now, so what can he do? He’ll answer some emails and binge-watch on YouTube, Reddit, and Blind until 2:00 a.m. before this day repeats all over again.
Conclusion
Aaaaand, that is the day of a typical mid-career senior data scientist. Now, close your eyes and go to sleep.
If you still can’t sleep, you can spend time on StrataScratch and read about data science or write code.
Don’t hold me accountable if you end up in data science! You can’t say I didn’t warn you — sorry, show you — how this Nate’s guy day looks.
Originally published at https://www.stratascratch.com.