Senior Data Scientist

Senior Data Scientist

myPOS

Sofia or Varna, Bulgaria

About the role:

As a Senior Data Scientist at myPOS, you will be the technical cornerstone of a lean, high-calibre team. Reporting directly to the Head of Data Science, you will independently own and deliver the most complex modelling programmes - setting the technical standard and ensuring that model work reaches production and creates measurable commercial value.

This is a role for a data scientist who has outgrown purely executing on defined problems and is ready to shape how problems are framed, lead architectural decisions, and hold end-to-end accountability for outcomes - not just model performance.

What you’ll do:

• Own the full lifecycle of complex modelling programmes across Customer & Commercial Intelligence: CLTV, Churn Prediction, Propensity to Buy, and Next Most Likely Product (NMLP);
• Architect multi-horizon churn models and build the churn intervention scoring layer that prioritises at-risk merchants for Account Management teams;
• Lead the NMLP engine - designing and productionizing a multi-output recommendation system that identifies the next product across myPOS's full catalogue;
• Take technical ownership of core fraud model components: transaction-level classifiers, merchant behaviour anomaly detectors, and new-account fraud scorers optimised for high-throughput, low-latency inference;
• Architect and own the Next Best Action (NBA) decisioning engine - a real-time system that selects the highest-expected-value action for each merchant at every interaction;
• Design and build production-grade agentic AI systems that automate high-value analytical and operational workflows;
• Define and execute experiment designs for online evaluation - A/B tests, uplift experiments, and bandits - and analyse results with statistical rigour;
• Set and enforce technical standards across the team: code quality, reproducibility, evaluation rigour, model documentation, and MLOps practices;
• Produce high-quality model documentation and present complex modelling work clearly to stakeholders across Sales, Marketing, Risk, Product, and Operations.

This role is perfect for you if you have:

• MSc or PhD in Computer Science, Statistics, Applied Mathematics, Econometrics or a related quantitative field (or equivalent commercial experience);
• 7+ years of applied data science and ML experience in a commercial environment, with a strong portfolio of models in production that drove measurable business outcomes;
• Expert Python for data science and ML engineering: pandas, scikit-learn, XGBoost/LightGBM, PyTorch or TensorFlow; clean, tested, modular code as a default;
• Deep expertise across the ML methodological spectrum: survival analysis, time-series and sequence modelling, uplift and causal inference, anomaly detection, and recommendation systems;
• Proven end-to-end ownership of at least three of: CLTV models, churn models, propensity models, fraud/risk models, recommendation or NBA systems - in a production commercial setting;
• Strong MLOps capability: feature stores, model registries, model serving infrastructure, drift monitoring, and CI/CD for ML pipelines;
• Deep SQL and data platform proficiency (GCP/BigQuery strongly preferred); experience with streaming architectures for real-time feature generation;
• Hands-on expertise building LLM-powered applications: RAG pipelines, tool-use agents, multi-agent orchestration, and agent evaluation frameworks;
• Strong experience with causal inference methods: uplift modelling, difference-in-differences, or instrumental variables;
• Excellent communication: able to present complex technical work to senior business stakeholders and write high-quality model documentation.

Nice to have:

• Experience in payments, fintech or financial services;
• Experience with reinforcement learning or contextual bandits for ranking and decisioning;
• Knowledge of graph neural networks for fraud or relationship modelling;
• Familiarity with AI governance frameworks (EU AI Act, SR 11-7);
• Published research or open-source ML contributions;
• Experience with real-time streaming inference (Flink, Spark Streaming).