Data Scientist
Extract actionable insight from complex data — combining statistics, programming, and domain knowledge to answer the questions that drive business and research decisions.
Data Scientists are hybrid professionals sitting at the intersection of statistics, programming, and domain expertise. They find meaningful patterns in large, messy datasets, build predictive models, design experiments, and communicate findings to decision-makers in ways that change strategy. While AI/ML Engineers focus on building and deploying ML systems at production scale, Data Scientists focus on the analytical and exploratory process — asking the right questions, cleaning and understanding data, building and interpreting models, and presenting insights with clarity. In practice, the boundary between Data Scientist and ML Engineer blurs depending on the organisation — at larger companies (Google, Meta, Grab) they are distinct roles; at smaller Sri Lankan companies they often overlap. In Sri Lanka, Data Scientist roles exist at Dialog Axiata (one of the largest telecom data operations in South Asia), Commercial Bank, Sampath Bank, John Keells Holdings, Hemas Holdings, PickMe, and a growing number of data-driven startups and technology companies. Internationally, Data Scientist has been consistently ranked among the top jobs globally since Harvard Business Review called it "the sexiest job of the 21st century" in 2012 — and demand has only grown since.
What a Data Scientist does daily
- Define analytical questions from business problems — translating "why are we losing customers?" into a churn prediction model specification
- Collect, clean, and prepare data — handling missing values, outliers, inconsistent formats, and joining data from multiple sources
- Conduct exploratory data analysis (EDA) — understanding distributions, correlations, trends, and anomalies in data before modelling
- Build statistical and ML models — regression, classification, clustering, time series forecasting, A/B test design and analysis
- Interpret and validate model results — understanding what the model is actually learning, checking for bias, and communicating confidence and limitations
- Build dashboards and visualisations — translating analytical results into charts, reports, and interactive dashboards for non-technical stakeholders
- Design and analyse experiments — A/B tests, multivariate tests, causal inference methods to distinguish correlation from causation
- Work with domain experts — finance, marketing, operations, clinical — to ensure models answer the right questions
- Communicate findings to executives and non-technical audiences — the ability to tell a clear, honest story with data is the most commercially valuable data science skill
- Stay current with analytical methods — Bayesian inference, causal ML, large-scale A/B testing, LLM-powered analytics
Step-by-Step Career Roadmap
- Build strong mathematics foundations — fractions, percentages, ratios, basic probability, and graphs are all data science fundamentals
- Develop curiosity about how data explains the world — sports statistics, weather patterns, election polling, economic charts
- Start using spreadsheets (Excel / Google Sheets) — sorting, filtering, basic charts; this is where data literacy begins
- Learn basic programming — Python via Code.org or Scratch; even very basic programming at this stage builds the right thinking habits
- Develop the habit of asking "why?" about statistics you see in the news or in school — this investigative habit is the core of data science
- Google Sheets / Excel project — analyse class test scores or sports data
- Scratch programming projects
- Explore a public dataset on Kaggle in a spreadsheet
- Sports statistics or election data analysis as a hobby
- Data science without mathematics is just data presentation — students who find maths tedious will struggle beyond entry-level analyst work; invest in maths foundations now