Data Scientist
Data Scientists model behavior and build systems that predict what happens next.
Don't just read it — play it
Clock in at 9am and make the calls a Data Scientist makes all day. Six decisions, real consequences, a score at 5pm.
What this role really is
Data Scientists use statistics and machine learning to answer questions that can't be answered by looking. You design experiments, build models, and quantify how confident anyone should be in the result. Expect more data plumbing and more meetings than the coursework suggests — and a real premium on knowing when a simple model is enough.
- Design experiments and A/B tests
- Build predictive models
- Clean and shape large datasets
- Quantify uncertainty for decision makers
- Ship models into the product
What it pays
- Entry level
- $100k–$140k
- Mid career
- $150k–$190k
- Senior
- $220k+
Typically 9:30–6
Majors that get you here
Entry roles: Data Scientist I · Research Intern · ML Analyst
Your 9-to-5, hour by hour
A realistic ordinary day — not the highlight reel.
9:30 AM
Experiment readout
The A/B test you launched two weeks ago has enough power. You check the guardrail metrics before you look at the win — the effect is real but half the size the team hoped.
10:30 AM
Feature engineering
You pull three months of behavioral data, build features for the churn model, and immediately find leakage in one of them.
12:30 PM
Model training
You train a baseline logistic regression before anything fancier, so you have an honest floor to beat. Gradient boosting adds four points of AUC.
2:00 PM
Working with engineering
You pair with a backend engineer on how the model will actually be served — latency budget, fallbacks, and what happens when a feature is missing.
4:00 PM
Explaining the result
You present the experiment readout to product leadership: what moved, what didn't, and the one thing you'd test next.
5:30 PM
Notebook cleanup
You commit the notebook, write up the methodology, and schedule the retraining job.
Great for you if
- You like open-ended quantitative problems
- You're comfortable saying 'we can't know that yet'
- You enjoy coding but not only coding
Probably not for you if
- You want fast, obvious wins every week
- You dislike ambiguity in what 'correct' means