Data Scientist, Greece
Role type: Full time
Location: Greece (fully remote)
Preferred start date: ASAP
Salary: €1400 - €3400 gross per month
ABOUT SATALIA
Satalia builds enterprise-grade AI systems for WPP and its FTSE 100 client base. Led by WPP Chief AI Officer Daniel Hulme, we run as a decentralised organisation where engineers and scientists own their domains end to end. Our data scientists are also at the core of WPP Research, conducting frontier research and developing prototypes that evolve into tools used in production across WPP.
THE ROLE
Our current work includes:
•Agentic systems. Developing long-horizon agents that plan, use tools, and collaborate. We also build the core infrastructure to ensure reliability and governance at scale, spanning context/memory, observability, traffic control, and continuous verification.
•Federated and privacy-preserving learning. Training shared models when the data cannot move: federated LLM fine-tuning across independent data owners, and aggregation strategies that hold up under heterogeneous datasets.
•Identity intelligence. Predictive and probabilistic modelling on large behavioural datasets: propensity and similarity models, and probabilistic record linkage across data sources.
•Geospatial intelligence. Spatial statistics, geospatial feature engineering, and forecasting and optimisation over location data, plus computer vision on imagery.
•Synthetic data. Generating realistic synthetic data with known ground truth to stress-test models and agents under controlled adversarial conditions.
You will own significant parts of these workstreams: choosing the modelling approach, building the training and evaluation infrastructure, and shipping to production by collaborating with engineers and domain experts across Satalia and WPP.
WHAT YOU'LL DO
•Own well-defined components of a project, from data and modelling to evaluation.
•Develop models, agents, and the evaluation harnesses that keep them safe and reliable.
•Work with engineers to take your work to production.
•Share your work clearly with the team and stakeholders, and support junior colleagues.
•Continuously upskill as the field evolves.
MUST HAVE
•2 to 5 years building ML/AI models in production.
•Strong modelling skills for both structured and unstructured data.
•Hands-on experience building LLM-powered systems.
•Good software engineering habits: Python, version control, testing, code review.
•Clear communication with both technical and non-technical stakeholders.
NICE TO HAVE
•Experience with temporal and geospatial data.
•Experience building and testing agents.
•Experience with federated learning, clean rooms, or other privacy-preserving methods.
•Background in marketing technology, ad tech, audience modelling, or media mix modelling.
•Publications or meaningful open-source contributions.
WHAT WE OFFER
•Healthcare benefits
•Remote working
•Flexible working hours
•Impactful projects
•Continuous learning and development
