It’s a lot of work and a lot of responsibility to be accountable for the company’s entire DS practice, and for people’s jobs and professional growth. It’s not for everyone! There are weeks when I don’t get to code at all.
You don’t have to be a manager or dept head to grow your career. You can be principal, lead, or senior data scientist, and you can accept some mentoring responsibilities but fight away the managerial and strategic ones.
(Sometimes it isn’t clear you have this choice in small companies. But in my experience, small companies also offer the most flexibility for crafting your role & growth.)
How did I grow into this role?
1. A part of it is what I touched on above. I was at a consulting company where the career promotion chart involved managerial responsibilities.
2. Another part of it is that I tend to spend time & “in-office political capital” to make sure my coworkers are doing okay.
Even before I was manager, if I had the attention of a senior person I’d mention if I thought one team was under-resourced.
That was NOT a great career move. But I knew that if I had influence inside a company that I wanted to use it to advocate for others.
I also am grateful to all the mentorship & career advice I had in academia and my career switch to data science. I also know firsthand how bad things get when your peers or seniors are inattentive or don’t care.
So I try to pay the good parts forward.
So I guess this second point is: make sure you *want* to manage.
That doesn’t mean just “manage work” but also mentor, advocate, lift up work, lose sleep for your coworkers.
3. As a consultant, I had the opportunity to work with executives at large companies.
I led and co-led projects that involved learning about their problems, and designing and iterating on solutions.
This gave me exp for the strategy and corporate project design part of the job.
4. I also feel that I should be candid about my privilege.
I’m a man with advanced degrees in STEM. My bosses liked to use the prestige of my degrees when introducing me to clients.
I can turn on, or often am unwittingly in, “confident physicist mode.”
I think those things amplified the actual execution of my work. It helped direct unique projects my way, and it helped the way people evaluated the outcomes.
These things help in some interview settings.
So when I went looking for my next job, I knew I was open to more managerial opportunities because of 2. Then 3 helped me convince people that I could lead strategy and project vision.
Okay, so what advice do I have for people who want to manage?
Let your boss know you want to manage.
Look for small opps first, like supervising an intern or being the onboarding buddy for a new hire.
Self-advocate to get opportunities to lead projects or be the point person.
And more generally, not everything you’re looking for in your career may come at the current job.
If you want to lead a team but your team has 3 people with no sign of growing, you may get some small leadership opportunities but need to look outside for the actual job title.
I should also remind you that I’ve been in data science for 4 years!
Yes, I have mentoring and leadership experience from academia—but I’ll be the first to acknowledge that I’m not the most seasoned veteran on this topic!
So grain of salt, only one opinion, etc!
• • •
Missing some Tweet in this thread? You can try to
force a refresh
Tuesdays are usually my most eventful days of the week, by design.
I have 9 meetings today, the earliest at 11 AM and the latest ending at 10:30 PM.
And you will ask: "Why, Taka, WHY in the name of reason and science would you do this to yourself?"
1. We have a regular call of all the department leads. That includes people in Korea, US ET + PT. Right now, we have it in the evening my time.
2. To minimize the number of evenings I'm in meetings, I put the rest of my meetings with Korea teams back-to-back with the above.
3. I also put cross-team meetings with stakeholders on Tuesdays, during the US day. These only happen every other week… but why also put them on Tuesdays!?
There are THREE reasons why I do this to myself (and my team, what a monstrous boss!).
Recent data: ~80% of technical roles in the biggest tech companies are held by men. wired.com/story/five-yea…
Further: ~92% of Fortune 500 CEOs are men, and I have yet to meet a female CTO.
Let me preface my answer to the question with Angela Davis’ famous quote: “In a racist society, it is not enough to be non-racist, we must be anti-racist.”
The point applies not only to racism, but to systemic inequity of all forms.
Users can read stories for free. For most stories, they'll hit a wall after several chapters. They can wait for the next chapter to unlock in an hour, or you can use in-app currency to read it right away.
Authors can publish their content, and we share the revenue it generates.
We have about 50 employees, about half of whom are in the US and half of whom are in Korea.
The DS team has 5 members, so we're a pretty big fraction of the company.
I’ll share a little bit here about my work day, throughout the day as I find time.
Here’s my “office,” a cramped corner of a bedroom where I’ve been doing all my work for almost a year—including interviewing for and hiring my teammates at the current job.
You will notice that my desk blocks the dresser door. It’s just as well—it’s not like I need blazers, suits, or ties these days!
Also, thank goodness for Zoom backgrounds! (I prefer astro images or Totoro for mine.)
My team starts the week with a 10AM sync on Monday.
How were our weekends? What are we working on this week, and what’s coming up? Anything holding us back? What are we looking forward to? Anyone taking days off?
In other words, you’re trying to predict whether they’ll like the movie.
The Qs might go: 1. Do you like sci-fi movies? 2. If yes, are you okay w movies w some violence? 3. If yes, do you like Jeff Goldblum? 4. If you don’t like Goldblum, do you like Laura Dern?
And so on.
After a while, you think you have a questionnaire that at the end, will be able to decide if you should recommend them to try Jurassic Park.
One class of machine learning algorithm, called decision trees, makes “questionnaires” kind of like this.