16 set - Pavia
Klarna
Overview In this role you will tackle technically challenging modelling problems in fintech by training transformer-based models on long sequences of real-world transactional events. You will design tokenisation schemes for numerical, categorical, and temporal features, shaping vocabulary size, sequence length, and information compounding. You’ll oversee the full model lifecycle from data prep to production, translating research decisions into systems that steer ML at Klarna.
You join a small, high-ownership team where your work directly impacts Klarna’s products. This is a mission-driven opportunity to shape scalable ML infrastructure in fintech.
Responsabilità
- Train transformer-based models on long sequences of transactions
- Design tokenisation schemes for heterogeneous features (numerical, categorical, temporal)
- Plan and optimize vocabulary size, sequence length, and information compounding
- Support end-to-end model lifecycle from data preparation to production serving
- Translate research decisions into production systems and ML direction for the company
- Collaborate in a small, high-ownership team to deliver impactful solutions
Requisiti fondamentali
- Deep understanding of transformer architectures and sequence modelling
- Experience designing tokenisation for heterogeneous features
- Proficiency in Python, PyTorch, SageMaker, and Airflow
- Experience owning the full model lifecycle from training to serving in production
- Ability to work on open-ended technical problems in a small team
- autonomy
- problem-solving
- curiosity
- transformer architectures
- sequence modelling
- tokenisation design for numerical, categorical, and temporal features
17 set - Bologna
Virtus Group SSD Fanball
17 set - Genova
Diablox
17 set - Portici
Call business
17 set - Ravenna
Ottica Dragoni