How to become a Data Scientist
Overview
Build the models and run the experiments that turn messy data into a product or policy decision you can defend.
BLS projects 34% growth (2024–34) for Data Scientists — the fastest of any role in this catalog and a top WEF growth area as companies put models in production, not just notebooks. Demand is broadening past pure research into applied ML, experimentation, and decision science.
What AI changes
What AI accelerates
First-pass model code, feature engineering drafts, hyperparameter search, summarising literature, and writing the first version of the analysis.
What stays human
Framing the problem, defending the causal claim, picking the right evaluation metric, deciding what would change your mind, and explaining it to non-technical leaders.
AI drafts model code, runs hyperparameter sweeps, and writes the first version of the analysis — but the scientist's edge is choosing the right question, defending the causal claim, and translating the model output into a decision the business will actually make. The judgement that justifies the model compounds; the routine parts of the job get faster.
Day to day
Partner with product or operations to scope a question, pull and shape the data, build and evaluate a model or experiment, write up findings, and present to stakeholders with the uncertainty spelled out.
Core skills
- Python (pandas, scikit-learn, PyTorch)
- Statistical reasoning & causal inference
- Experiment design
- SQL & data pipelines
- Communication of model results
Tools
- Python (pandas, scikit-learn, PyTorch)
- SQL
- Jupyter / notebooks
- Experiment platforms (Optimizely, in-house)
- Tableau/Looker for downstream reporting
How to get in
Entry routes
- From a data analyst role with a stats/ML upskilling path
- From a quantitative PhD or research role
- From a software engineering role with applied ML focus
Seniority ladder
| Level | Title | Experience | Focus | Salary |
|---|---|---|---|---|
| Entry | Junior Data Scientist | 0–2 yrs | Building models with supervision, learning the business | Entry of the US band, below the role median |
| Mid | Data Scientist | 2–4 yrs | Owning models end-to-end, partnering with product | Around the role median |
| Senior/Lead | Senior Data Scientist | 4–7 yrs | Leading the science on a product area, mentoring | Upper end of the US band |
| Director | Director of Data Science | 7+ yrs | Science strategy, team leadership, cross-function impact | Above the IC band, with a management premium |
Where it can lead
Progresses to
- Senior Data Scientist
- Director of Data Science
- machine-learning-engineer
- Head of Data
Pivots to
- machine-learning-engineer
- analytics-engineer
- data-engineer
- product-manager
Pay (US)
USD 95,000
USD 112,590
USD 175,000
Outlook
Data Scientists is the fastest-growing US occupation BLS tracks: 34% projected growth 2024–34, well above the 3% all-occupation average.
Prove it
Churn Driver Analysis Memo
A/B Test Write-Up From a Public Experiment
Pricing Experiment Design Memo
Power & Sample-Size Justification Memo
Survival Analysis Reproduction
Interview prep
Walk me through a model you built from exploration to production.
How do you know when a model is good enough to ship?
Your path into Data Scientist
See how your experience lines up — skill gaps, salary fit, and a personalised seniority match. No invented claims, just your real career mapped against this role.
Unlock all 10 career paths + deep reports
See full fit breakdowns, skill-gap maps, proof-project ideas, and salary outlooks for every path.