Data Scientist – Credit Risk and Fraud

At Madiff, we connect top IT talent with cutting-edge companies through remote Agile teams and nearshore outsourcing. With operations in Poland, Portugal, France, and the UK, we specialize in AI, DevOps, and Cybersecurity – supporting industries like Telecom, Banking, and Hi-Tech.

Are you motivated by building production-grade machine learning models that directly influence credit approvals and fraud prevention in live banking systems? Do you want your work to operate at scale, under strict regulatory and performance requirements, and truly impact real financial decisions across international markets?

We’re hiring a Data Scientist – Credit Risk and Fraud to join a global programme delivering a core enterprise decision intelligence platform for a major banking group. The platform combines large-scale structured data processing, advanced ML models, and GenAI orchestration layers to support real-time and batch decisioning in live banking environments.

You’ll work on models that generate risk signals consumed directly by lending engines, fraud detection systems, and downstream decision workflows – within a production system operating at significant scale.

Key responsibilities:

Design, build, and maintain credit risk and fraud detection models influencing real customer decisions.

Perform advanced feature engineering on large-scale structured financial datasets.

Train, validate, and optimise ML models (e.g. gradient boosting) with strong focus on robustness and interpretability.

Monitor and improve models in production, ensuring performance, stability, and regulatory alignment.

Integrate model outputs into LangChain and LangGraph-based orchestration pipelines supporting decision logic and escalation workflows.

Collaborate closely with ML engineers on deployment, monitoring, and experiment tracking using MLflow.

Support regulatory documentation, model governance, and explainability processes in a controlled banking environment.

Work in a cross-functional squad (6–9 people) including data scientists, ML engineers, and a dedicated tech lead, with active involvement of risk and finance stakeholders.

What we are looking for:

Strong experience as a Data Scientist working with production ML models.

Very good Python skills and practical experience with pandas, NumPy, and scikit-learn.

Hands-on experience with XGBoost or LightGBM.

Strong SQL skills and experience with large-scale data processing (PySpark).

Experience with MLflow for experiment tracking and model lifecycle management.

Understanding of model validation, monitoring, and explainability in regulated environments.

Experience or familiarity with LangChain and LangGraph for orchestration of analytical outputs.

Strong communication skills and ability to collaborate with technical and business stakeholders.

Fluent English (project language).

Availability to work in standard Polish business hours.

What we offer:

Opportunity to work on a live, enterprise-scale decision intelligence platform in the global banking domain.

Real production impact – your models directly influence credit approvals and fraud prevention.

Modern ML ecosystem with structured data processing, orchestration layers, and MLOps practices.

Cross-functional collaboration with data scientists, ML engineers, tech leads, and risk stakeholders.

Agile ways of working and ownership over model quality, performance, and production readiness.

Recruitment process:

Initial screening call with Madiff.

Technical interview focused on modelling and data handling.

Client interview covering project fit and collaboration.

Final decision within 5 days maximum from the first interview.