Lead Data Scientist

Xebia is a global tech company with over 1,000 experts delivering work across cloud, data, and software. Originating from two Polish companies, PGS Software and GetInData, Xebia serves clients in fintech, e-commerce, aviation, logistics, media, and fashion, including McLaren, Aviva, Deloitte, Spotify, Disney, ING, UPS, Tesco, Truecaller, AllSaints, Volotea, Schmitz Cargobull, Allegro, and InPost. The company uses modern open-source stacks and is a trusted partner of Databricks, dbt, Snowflake, Azure, GCP, and AWS.

• Solid statistics background including regression, GAMs, mixed/random effects, link functions, robust estimation, and outlier handling
• Proficiency in Python and data science stack: NumPy, Pandas, statsmodels, SciPy, scikit-learn
• Experience building and maintaining production APIs with FastAPI and Pydantic
• Comfortable working with PostgreSQL and SQLAlchemy
• Familiarity with containerized environments such as Docker, Kubernetes, and GCP
• Ability to translate domain requirements into quantitative solutions and communicate trade-offs
• Good command of English (spoken and written)
• Familiarity with basic statistical concepts like Bayes’ rule, linear regression, and maximum likelihood estimation
• Practical experience using AI-powered assistants (e.g., Claude Code, GitHub Copilot, Cursor) to improve productivity, quality, or decision-making in software delivery

Nice to have:

• Experience with geospatial data and libraries (GeoPandas, Shapely, H3, GeoAlchemy2)
• Knowledge of GAM libraries (PyGAM), JAX, or TensorFlow Probability
• Experience with task queues and async workflows (Celery, Redis)
• Familiarity with observability tooling (OpenTelemetry)
• Experience with ML pipeline frameworks (Kedro)
• Data validation and property-based testing (Pandera, Hypothesis, TestContainers)
• R integration (rpy2)
• LLM integrations (Google Gemini or similar)
• Frontend awareness (React, TypeScript)
• Knowledge of real estate data, valuation methodology, or appraisal workflows

• Designing and developing statistical models for property price adjustments across time, location, quality, and condition
• Building spatial algorithms such as adaptive heatmaps, geographic clustering, and polygon-based property search to capture local market dynamics
• Implementing comparable property recommendation with feature engineering across different property types
• Developing market analysis pipelines with diagnostics including trend fitting, outlier detection, and goodness-of-fit metrics
• Integrating LLM-based classification services for document and property analysis
• Exposing model outputs through production API endpoints and collaborating with frontend engineers on data contracts
• Debugging models in production addressing edge cases, numerical issues, and data quality problems

Work from the European Union region and a work permit are required.