- Munich, Bavaria, Germany
- Hybrid
- Full-time · Staff
- 5+ years
Build and bring into production state-of-the-art probabilistic models that power millions of pricing and offer decisions at SIXT, shaping both what we charge and what we show.
Responsibilities
- Architect and prototype advanced regression models with a principled Bayesian approach
- Design and productionise personalisation and recommendation systems
- Apply techniques like SVI, MCMC, and importance sampling for robust decision-making
- Build and maintain reproducible pipelines for feature engineering, label generation, and automated data validation
- Deploy models via FastAPI, Docker, or Kubeflow and establish real-time monitoring
Requirements
- Nice to have: Two-tower retrieval
- Nice to have: Embedding-based candidate generation
- Nice to have: Learning-to-rank
- Nice to have: Sequential models
- Nice to have: Contextual bandits