Pershiy Ukrayinskiy Mizhnarodniy Bank, AT / PUMB
Qualifications and experience.Required2+ years of machine learning / data science experience, preferably in financial sector or fintech.Experience in building anti-fraud models or fraud detection systems.Proficient command of Python, experience with basic ML frameworks: scikit-learn, XGBoost, LightGBM, CatBoost.Knowledge of algorithms for classification, regression, clustering, anomaly detection.Experience with SQL — data extraction, aggregation, joins, query optimization.Practical understanding
Qualifications and experience.
Required
- 2+ years of machine learning / data science experience, preferably in financial sector or fintech.
- Experience in building anti-fraud models or fraud detection systems.
- Proficient command of Python, experience with basic ML frameworks: scikit-learn, XGBoost, LightGBM, CatBoost.
- Knowledge of algorithms for classification, regression, clustering, anomaly detection.
- Experience with SQL — data extraction, aggregation, joins, query optimization.
- Practical understanding of model evaluation metrics: ROC-AUC, Precision/Recall, Gini, KS, F1.
- Experience building models based on behavioral features, time-series data and transactional patterns.
- Skills for working with APIs to interact with a production environment in real time.
- Understanding CI/CD principles, experience working with Git. Experience in launching and monitoring models in production, processing large volumes of data in real time.
Desired
- Experience in building rule-based + ML hybrid models in the field antifraud.
- Knowledge of financial products (cards, accounts, loans, payments) and typical fraud schemes.
- Working with graph structures (graph-based fraud detection, network analysis).
- TensorFlow, PyTorch — as a bonus in deep learning or NLP tasks.
Functional responsibilitiesDevelopment and improvement of fraudulent transactions in real life time.Preparation of large arrays of transactional, behavioral and additional data for training models.Performance of data profiling, cleaning, transformation, physical engineering.Identification of new fraud patterns based on historical data and trends.Testing, validation and optimization of models using relevant quality metrics.Integration of models in production environment, maintenance, stability control, updates.Scaling of antifraud solutions for highly loaded systems with a large number of transactions.Collaboration with teams of analysts, developers and product managers to define business requirements and integrate models into business processes.Participation in projects to introduce innovative approaches to financial security, with the ability to launch initiatives at the level of the entire antifraud strategy.Why PUMB?
- A strong antifraud team with a focus on building intof electronic anti-fraud systems that works with modern tools and approaches.
- Working in one of the most technologically advanced banks of Ukraine, with a real impact on the security of millions of customers. The possibility of implementing ML models in a real business process.
- Wide access to data, high autonomy in decision-making, support from management.
- Hybrid or remote work format, competitive conditions hiring
- Strong culture of interaction, transparency, openness to innovation and personal development.
- Participation in projects at the intersection of Data Science / Antifraud / Digital Banking — areas that are actively transforming at the global level.
