A fictional person: the names, employers and numbers are invented.
Daniel Okafor
Data Scientist · machine learning, experimentation, Python
Edinburgh, UKdaniel.okafor@example.com
Summary
Data scientist with five years' experience taking machine-learning models from notebook to production in lending and online retail. Comfortable owning the whole loop: framing the question, building features, shipping the model with engineers and proving its value with a proper experiment. Latest credit-risk model cut bad-debt losses by £1.3 million a year.
Experience
Data Scientist · Northbridge Lending, Edinburgh
2022 – Present
Rebuilt the credit-risk model with gradient boosting and new open-banking features; Gini rose from 0.52 to 0.61 and bad-debt losses fell £1.3m a year.
Deployed models as versioned services with MLflow and Docker, adding drift monitoring that flags problems within a day instead of a quarter.
Designed the firm's first A/B testing guidelines, including power calculations, now used for every pricing and onboarding change.
Worked with the model-risk team to document fairness checks across age and postcode groups ahead of an external audit.
Junior Data Scientist · Lothian Retail Group, Edinburgh
2020 – 2022
Built a product-recommendation model for the website that raised average basket value by 6% in a four-week controlled test.
Forecast weekly demand for 3,000 products with Prophet and LightGBM, reducing fresh-food waste by about 11%.
Ran a churn analysis that showed delivery-slot failures drove a third of cancellations, leading to a new slot-booking rule.
Education
MSc Data Science · University of Edinburgh
2019 – 2020
Distinction; thesis on uncertainty estimates for demand forecasting.
A/B testing, power analysis, causal inference, Bayesian methods
Engineering
Python, SQL, MLflow, Docker, Airflow, dbt
Cloud
Google Cloud (BigQuery, Vertex AI), Git
Projects
Open demand-forecasting benchmark
2021
Co-wrote a public notebook comparing six forecasting methods on open retail data, later used as teaching material in a university course.
Languages
English — Native
French — Intermediate (B1)
Three summary options
Start from the one closest to where you are now.
Recent MSc graduate
Data science MSc graduate with a strong statistics background and a summer placement at an insurer. Built a claims-fraud classifier that flagged 30% more suspicious claims at the same review workload. Writes clean, tested Python and wants to learn how models live in production.
Lead data scientist
Lead data scientist with nine years in marketplaces and payments. Manages four data scientists, sets the experimentation standards for 20 product teams and owns the pricing models behind £80m of annual revenue. Known for turning vague questions into testable ones.
From academia
Computational biologist with a PhD and four years of postdoctoral research, now moving into industry data science. Built Bayesian models on datasets of two million records, published six peer-reviewed papers and taught statistics to 150 students a year.
Keywords to include
Use the ones that are true for you, inside real sentences.
Machine learning
Python
SQL
A/B testing
Statistical modelling
scikit-learn
Feature engineering
MLOps
Forecasting
Data visualisation
What recruiters look for
Tips for this role, with the kind of numbers that count.
Prove the model reached production
Many CVs stop at accuracy scores from a notebook. Say where the model runs, who uses it and how you monitor it. "Deployed with MLflow, with drift alerts within a day" separates you from people who never shipped.
Translate model metrics into money
Hiring managers outside your team may not know what a Gini of 0.61 means. Pair the technical metric with the business one: losses avoided, revenue gained, hours saved or waste reduced.
Show you can run an experiment
Controlled tests are how data scientists earn trust. Mention test design, sample size or power, and the result, for example "basket value up 6% in a four-week test". Stating the duration makes the claim more credible.
Group your skills by job
A flat list of 30 libraries is hard to read. Group them into modelling, experimentation, engineering and cloud, as in the example, so a reviewer sees in seconds whether you cover the parts of the role they need.
Mention responsible use of data
In lending, health and insurance, fairness, privacy and explainability matter. A bullet about fairness checks, model documentation or working with a risk team shows you understand the rules your models live under.
Keep academic work relevant
Your thesis or papers belong on the CV if they show a skill the job needs, such as forecasting or causal inference. Summarise each in one line with the method and the outcome rather than the full title.
Make this CV yours
Answer one question at a time; the assistant fills each section and never invents numbers.
Should I list Kaggle competitions on a data scientist CV?
Only strong results, and only if you have little industry work to show. A top 5% finish shows modelling skill, but employers value deployed models and run experiments more. As your career grows, swap competitions for work projects with measured business impact.
How technical should a data scientist's summary be?
Lightly technical. Name your main area, such as forecasting or credit risk, the core tools and one business result. The summary is often read by a recruiter first, so save algorithm names and metric details for the experience section.
Do I need a PhD to apply for data scientist roles?
Most industry roles do not ask for one; an MSc or a strong numerate degree plus experience is common. If you have a PhD, present it as research experience with clear skills. If you don't, let production work and experiments carry the weight.
