Data Scientist

myPOS

Sofia or Varna, Bulgaria

About the role:

myPOS is building a high-impact Data Science function to power the intelligence layer of one of Europe’s fastest-growing payment and commerce platforms. As a Data Scientist, you will contribute to a focused team working across a rich portfolio of models that drive smarter decisions in Sales, Marketing, Risk, Operations, Product and Technology.

You will move fluidly across problem types: from customer lifetime value and churn modelling to fraud scoring and agentic AI workflows.

What you’ll do:

Build and maintain ML models across the core portfolio: CLTV, churn prediction, propensity to buy, and Next Most Likely Product (NMLP).
Develop fraud detection models including transaction-level classifiers, merchant behaviour anomaly detectors, and new-account risk scorers.
Contribute scored model outputs to the Next Best Action (NBA) decisioning layer that selects the optimal action for each merchant across Sales, Marketing, and in-product touchpoints.
Support A/B experiments, uplift tests, and multi-armed bandit evaluations to measure the incremental impact of model-driven interventions.
Design and implement end-to-end ML pipelines - from data ingestion and feature engineering through to model training, evaluation, and deployment.
Monitor deployed models in production: detect performance degradation, data drift, and data quality issues; iterate and document changes proactively.
Collaborate with business teams across Sales, Marketing, Risk, Operations, and Product to translate business problems into well-defined data science solutions.
Run rigorous experiments and communicate findings clearly to both technical and non-technical stakeholders.
Contribute to LLM-powered agentic workflows using tool-use patterns (RAG, function calling, memory) and frameworks such as LangChain or LlamaIndex.
Contribute to team documentation: model cards, methodology write-ups, and internal playbooks that help the team scale its practices.

This role is perfect for you if you have:

3–5 years of hands-on applied data science, machine learning or statistical modelling experience in a commercial setting, with models shipped and measured in production.
Strong proficiency in Python for data science: pandas, numpy, scikit-learn, XGBoost / LightGBM, and at least one deep learning framework (PyTorch or TensorFlow).
Solid grounding in supervised and unsupervised learning: classification, regression, clustering, survival analysis, and time-series modelling.
Demonstrable experience building at least one of: CLTV, churn, fraud detection, propensity, or uplift models in a production environment.
Comfort working with large-scale structured and semi-structured data; proficient in SQL and cloud data warehouses - GCP and BigQuery strongly preferred.
Familiarity with ML experiment tracking platforms (MLflow, Weights & Biases) and model serving patterns (REST APIs, batch inference pipelines).
Working knowledge of LLM APIs (OpenAI, Anthropic, etc.) and at least one agentic AI framework (LangChain, LlamaIndex, AutoGen, or similar).
Understanding of responsible AI: fairness assessment, model explainability methods (SHAP, LIME), bias detection and mitigation strategies.
Clear communication - able to distil statistical findings into actionable insights for both technical peers and business stakeholders.