7 lessons every junior data scientist must learn to stop being just “the data person” and start becoming the teammate everyone trusts
(Even if you're shy)
Hey there!
Do you feel like you’re just coding into the void?
Like, technically you’re doing “data science”—but somehow, it doesn’t feel like it?
Your work barely gets used.
Stakeholders don’t loop you in early.
And honestly, you start wondering if any of it even matters.
Yeah... I’ve been there.
And I know that I am not the only one, because here is a message that I received…
Back when I started, I was laser-focused on doing the work right.
But here’s the kicker—I never stopped to ask if I was doing the right work.
So I ended up with beautiful models that no one used, hours cleaning data no one asked for, and weeks of silence followed by, “Yeah, this isn’t quite what we needed.”
So today, I want to share 7 hard-earned lessons that helped me stop being just a quiet task-taker and start becoming someone who actually drives impact, is a trusted team member and gets promoted.
Let’s dive in.
1. Ask “Why” before you write a single line of code.
Okay, real talk—when I first started, I’d open up my notebook and just go.
You know the drill:
Tweak the model. Tune some hyperparameters. Add that one last metric just for good measure.
But here’s what I learned (the hard way):
If you don’t pause and ask why you’re even doing the thing… you might be building something that literally no one needs.
Like, at all.
So now, I force myself to stop and ask a few simple questions:
“What decision is this helping us make?”
“Who’s gonna use this?”
“Why now?”
You do that, and suddenly your work starts to matter more. Not just to you, but to the team.
2. Focus on inputs, not just outcomes.
It’s so tempting to chase the flashy stuff when you're just getting started.
I mean, who doesn’t want to hit 95% accuracy or drop a dashboard that looks like it belongs in a product launch deck?
But here’s the thing I had to unlearn:
The best results? They don’t come from chasing the output.
They come from obsessing over the inputs.
That means asking:
“Did we even frame the problem right?”
“Are we solving the right thing?”
“Is the data even legit?”
If you nail the setup, the shiny stuff takes care of itself. Promise.
3. Take extreme ownership—even when you think it’s not your job.
This one changed everything for me.
There was a time when I’d shrug things off.
Like, “Well, no one gave me clear specs,” or “Pretty sure that’s the PM’s thing, not mine.”
But honestly? That mindset keeps you small.
The people who grow fast—who become really valuable—they don’t play the blame game.
They say, “Alright, this is messy... but I’ll figure it out.”
And then they go clean the data. Or chase down the missing context. Or fix the broken dashboard even if they didn’t build it.
That kind of ownership? It builds trust. And trust? That’s your rocket fuel.
4. Create feedback loops that are stupidly short.
Oof, this one still stings.
I once spent two full weeks building out a whole solution—slides, visuals, everything.
Showed it to the PM and guess what?
“Yeah… this isn’t really what we needed.”
Crushed me.
So now? I don’t wait. I send rough drafts. Sketches. Looms.
I’ll literally ask, “Hey, is this even close to what you had in mind?” within the first couple hours.
It’s not about showing off polish—it’s about making sure you’re even in the right ballpark.
5. Communicate with radical clarity—especially in writing.
You ever write an update that makes total sense to you... and then no one gets it?
Been there. Too many times.
The trap is either being way too technical or way too vague.
What works better? Just say it like it is:
Here’s what I’m doing
Here’s why it matters
Here’s what’s in the way
That’s it. Keep it simple. Say the thing.
Your updates aren’t about sounding smart—they’re about getting people on the same page.
6. Overcommunicate—even when you think you already did.
This one always surprises people.
You might be thinking, “But I already explained that in the Jira ticket!”
Sure. You did. But:
PMs didn’t read your code
Your boss forgot your last update
Nobody is sitting around refreshing your dashboard
So yeah—say it again. And again. And maybe one more time, just to be safe.
Seriously:
Drop a weekly Slack update
Record a 1-min Loom
Write a 3-bullet summary
Repetition isn’t annoying. It’s how people actually hear you.
7. Shift from delivering output to delivering impact.
Okay, this is the part where it all starts to click.
When you put everything above into practice—asking “why”, owning your work, tightening feedback loops—you start to feel this shift.
You’re not just writing code anymore. You’re driving things forward.
You’ll notice it when:
Your analysis gets quoted in meetings
PMs start looping you in early
People start asking for your take
And let me tell you—when your work starts shaping real decisions? That’s when this job gets fun.
So... where should you start?
Look, don’t try to do all 7 at once. You’ll burn out.
Just pick one. Any one. Like:
Pause and ask “why” before your next ticket
Share something rough and ask for quick feedback
Send a Slack message saying what you’re working on
You don’t need to add more code to your repo.
You need to build more connection to the work—and the people around it.
That’s the shift.
And you’ve totally got this. 💪
Until next time,