Production ML Engineer — Credit Modeling Pipelines

Klarna; Milano; Ibrido; EUR 90.000 - 130.000; Tempo pieno

Descrizione del lavoro

Klarna is seeking a Lead Engineer for the Credit Modeling Pipeline to transform consumer credit underwriting models from experiments into production-ready systems. You will write production-grade Python, build the ML infrastructure, and deploy models using AWS SageMaker, ensuring reliability in production.

You will work with a cross‑functional team across Stockholm and Warsaw, with a split base in Milan or Stockholm per location, and help scale Klarna’s in‑house data science capability as the

Competenze

• Production Python for ML
• Takes ML models to production and owns them
• Hands-on with tree-based models
• Deploys ML workloads on cloud platforms (AWS or similar)
• Follows full SDLC for ML code and workflows
• Collaborates with data scientists to build reliable pipelines
• Communicates effectively in English (spoken & written)

Mansioni

• Write Python to train credit underwriting models and prepare them for production.
• Build and maintain infrastructure for the ML pipeline, from feature computation to retraining and monitoring.
• Deploy models into production using AWS SageMaker and keep them running.
• Troubleshoot and fix pipeline issues end to end.
• Scale Klarna's in-house data science capability as credit risk and fraud teams grow.

Conoscenze

Python

ML deployment

AWS SageMaker

CI/CD for ML

Production-grade code

Tree-based models

Strumenti

AWS SageMaker

Descrizione del lavoro

Klarna is seeking a Lead Engineer for the Credit Modeling Pipeline to transform consumer credit underwriting models from experiments into production-ready systems. You will write production-grade Python, build the ML infrastructure, and deploy models using AWS SageMaker, ensuring reliability in production.

You will work with a cross‑functional team across Stockholm and Warsaw, with a split base in Milan or Stockholm per location, and help scale Klarna’s in‑house data science capability as the