The Generalist
Master of Many Trades — handles the full spectrum from data collection through deployment. Essential in startups where breadth trumps depth.
← Fundamentals
1.3 · Fundamentals
Types of Data Scientists
Six distinct specialisations have emerged as data science matured — each bringing unique value, each requiring a different blend of skills. Effective organisations build complementary teams, not impossible unicorns.
The Taxonomy
Six specialisations
Common patterns rather than rigid categories — real practitioners often blend characteristics, and career paths frequently cross boundaries.
The Analyst
Bridge Between Data and Business — specialises in BI, reporting, and insight generation. Communication and business acumen are the differentiators.
The ML Engineer
From Models to Production — bridges data scientists and software engineers, ensuring models deliver value reliably at scale.
The Researcher
Pushing Boundaries — develops novel algorithms, publishes at NeurIPS/ICML/ICLR, and expands the toolkit available to all practitioners.
The Product Scientist
Data Science for Growth — specialises in A/B testing, user behaviour analysis, and growth optimisation at consumer tech companies.
The Data Engineer
Foundation Builder — builds and maintains systems that collect, store, transform, and serve data. Without them, nothing else works.
Specialisation 1
The Generalist
Most common in startups and small teams. Breadth enables complete solutions; depth may lag specialist teams.
Skills Profile
Core strengths: Python & R proficiency · SQL competency · basic ML algorithms · data visualisation · cloud platform familiarity · business communication. Breadth without sacrificing foundational depth — capable of working independently on moderate-complexity problems.
A Typical Week
- Stakeholder meetings to scope new projects
- SQL queries against production databases
- Exploratory analysis in Python notebooks
- Training and evaluating models
- Containerising applications for deployment
- Monitoring production systems for issues
Salary & Career Path
- Entry: £40k–£55k
- Mid: £55k–£85k
- Senior: £85k–£120k
Academic backgrounds: physics, economics, engineering, computational sciences. Senior progression leads to team leadership or specialist transition.
Specialisation 2
The Analyst
Descriptive analytics that explain what happened and why. Primary value lies in translating data into business understanding.
Skills Profile
Core strengths: SQL · Tableau · Power BI · stakeholder communication · executive presentations · statistical interpretation (p-values, confidence intervals). Communication skills distinguish excellent analysts from merely competent ones — lead with conclusions, choose visualisations that make insights immediately apparent.
Daily Activities
- Translate vague business questions into specific analytical problems
- Query data warehouses with SQL
- Build dashboards in Tableau or Power BI
- Prepare executive presentations — lead with conclusions
- Conduct ad-hoc deep-dives on unexpected patterns
- Build trust through consistent delivery and honest uncertainty
What Sets Them Apart
- Translate p-values into revenue implications
- Structure narratives for executives — strategy not statistics
- Visualisations that make patterns immediately apparent
- Sometimes build simple predictive models
Salary ranges mirror Generalist trajectory; senior analysts at data-mature organisations can command £80k+.
Specialisation 3
The ML Engineer
Infrastructure, pipelines, and systems — ensuring models deliver value reliably rather than remaining interesting experiments.
Skills Profile
Core strengths: production-quality Python · Docker · Kubernetes · CI/CD practices · Prometheus & Datadog for monitoring · cloud platform expertise. Software engineering and infrastructure over pure ML theory.
What They Build
- Data pipelines ensuring reliable training data flow
- Training pipelines with experiment tracking & model versioning
- Serving infrastructure — thousands of predictions per second
- Feature stores ensuring training/inference consistency
- Monitoring systems detecting model degradation
- Fraud detection and similar high-throughput production systems
Salary & Career Path
- Mid: £60k–£90k
- Senior: £90k–£140k
Bridges data scientists and production engineering. Strong demand driven by the gap between model building and reliable deployment at scale.
Specialisation 4
The Researcher
Advances the state of the art — primarily in academic institutions, industry labs (Google DeepMind, Meta AI Research), and specialised AI organisations.
Research Impact
Impact manifests differently from applied data science. A researcher developing a new attention mechanism may not immediately see business application — but years later that mechanism might power products serving millions. Researchers expand collective knowledge, enabling future practitioners to solve currently intractable problems.
Research Activities
- Read academic papers to understand knowledge frontiers
- Design controlled experiments testing algorithm behaviour
- Derive mathematical proofs and convergence guarantees
- Develop novel algorithms with theoretical foundations
- Write & submit papers to NeurIPS, ICML, and ICLR
- Implement reference implementations of new approaches
Salary & Entry Requirements
- Mid: £80k–£120k
- Senior: £120k–£200k
Doctoral degree in computer science, statistics, mathematics, or related field typically required. Industry researchers (Google DeepMind, Meta AI Research) substantially out-earn academic counterparts.
Specialisation 5
The Product Scientist
Emerged from the intersection of data science and product management. Prominent at consumer technology companies where millions of users enable rapid experimentation.
Skills Profile
Core strengths: statistical experimental design · causal inference · hypothesis testing · statistical power · multiple testing corrections · confidence intervals. Must balance statistical rigour with product velocity — high power might require months; product teams demand faster iteration.
Daily Activities
- Design A/B tests for product changes (engagement, retention, revenue)
- Implement statistical analysis ensuring adequate experiment duration
- Apply multiple testing corrections to avoid false positives
- Build dashboards tracking product health metrics
- Segment users and analyse behaviour across cohorts
- Quantify the causal impact of product decisions
Salary & Career Path
- Mid: £55k–£85k
- Senior: £85k–£130k
Partners closely with product managers and engineers. Growth optimisation and user behaviour analysis are the primary outputs.
Specialisation 6
The Data Engineer
Increasingly a core team member — quality data infrastructure is a prerequisite for effective analytics and modelling. Without data engineers, data scientists wait.
Skills Profile
Core strengths: Python & Scala · SQL mastery · Airflow & Prefect for orchestration · Spark for distributed computing · cloud platform expertise. Combines software engineering with deep data system knowledge.
What They Build & Maintain
- Extraction pipelines from APIs, databases, event streams & file systems
- Centralised data lakes and warehouses
- Transformation pipelines producing clean, derived datasets
- Data quality monitoring — missing values, schema changes, distribution shifts
- Access controls enabling secure self-service analytics
- Query optimisation for interactive analyst exploration
Salary & Career Path
- Mid: £50k–£80k
- Senior: £80k–£120k
A growing specialisation as organisations recognise that data infrastructure is the foundation on which all other data work depends.
Compensation
UK salary ranges by specialisation
Senior maximum · annual gross. Researchers command a premium reflecting doctoral requirements and research lab demand.
Team Design
Building effective teams
The composition that works depends on organisational size, data science maturity, and business objectives — not on finding unicorns who excel at everything.
Startup ~3 people
- 2 × Generalist data scientists
- 1 × Data engineer
Generalists handle everything adequately; the data engineer keeps infrastructure from becoming a bottleneck.
Mid-size ~10 people
- Several analysts
- A few data scientists
- 2 × ML engineers
- 2 × Data engineers
Specialisation becomes both possible and beneficial as complexity grows.
Large Tech 100s
- Product scientists embedded with product teams
- ML engineers supporting production systems
- Researchers exploring emerging techniques
All specialisation types working together, with clear roles and handoffs.
Key Insight on Teams — Hire specialists who complement each other rather than holding out for rare individuals who excel at statistics, programming, communication, and business strategy simultaneously. An analyst who communicates brilliantly but has moderate programming skills pairs well with a data scientist who codes elegantly but struggles with presentation. The team's collective capability matters more than any individual's completeness. Effective data science requires diverse capabilities distributed across specialists.
Key Takeaways
- Data science encompasses diverse specialisations beyond the generic "data scientist" title
- Generalists suit small teams; specialists enable scale and sophistication
- Analysts bridge data and business; ML Engineers bridge models and production
- Researchers push boundaries; Product Scientists optimise growth
- Data Engineers provide the foundation enabling all other data work
- Effective teams combine multiple types rather than seeking unicorns
- Career paths accommodate both specialists deepening expertise and generalists broadening skills