Senior Data Scientist

Twoday is one of the leading digital transformation partners in Northern Europe with a global presence. With approximately 3,000 technologists, we collaborate with the most admired private and public organizations to deliver cutting-edge digital solutions. Our deep industry expertise includes Data & AI, software development, digital experiences, and business applications. Operating across the Nordics and Lithuania, our team serves over 8,000 customers and supporting their digital transformation journeys.  We are looking for an experienced Senior Data Scientist to join our AI team! Most organizations have no shortage of ideas for what to model. Far fewer can turn a vaguely stated business problem into a well-posed prediction task, pick the algorithmic approach that fits it, and defend the result when someone asks how the number was produced. That gap is where you come in. You will do the analytical work that turns a customer's business problem into a model that holds up, and then into something they actually use, alongside data engineers, AI/ML engineers, cloud architects, and software developers on some of the largest Data & AI teams in the Nordics. This is a hands-on role. You will spend most of your time in data and in code, with experienced colleagues around you and room to take ownership of the analyses and models you know best. What are we looking for? You are a scientist and a builder. Given a business problem stated in a customer's own words, you can work out what should be predicted, at what granularity, against which target, and what a good result would even look like. You are rigorous about how you validate a result, and you have learned that the hard part is rarely the algorithm itself. We expect you to bring: Hands-on data science experience, including at least one or two models you have taken past the prototype stage and to production. Experience translating a business objective into a formal AI/ML problem: defining the target variable and unit of analysis, agreeing what success means with the people who will use the model, and saying early when the data will not support the question. Strong predictive modelling craft: classification, regression, time series and generative AI Judgement about algorithmic choice: when a well-specified baseline or classical statistical model is enough or when a heavier method earns its cost and when the problem is not a prediction problem at all. Solid applied statistics: honest validation, uncertainty, and the ability to explain what a result does and does not support. Strong software fundamentals to write code others can build on Experience on Azure: Azure Machine Learning, Databricks, or Microsoft Fabric (Experience from AWS or GCP transfers, and we will not hold it against you.) Experience applying generative AI where it genuinely fits, including RAG architectures and the evaluation of prompts and models. An awareness of what governance means in practice: reproducibility,…