Your Job:
This PhD project develops a Bayesian inference framework for hybrid model- and data-driven modeling of metabolism, with a particular focus on handling model misspecification. By combining Bayesian computational statistics, differentiable programming, and high-performance computing, the project aims to deliver robust, interpretable, and scalable methods for metabolic flux analysis.
You will:
1. Design hierarchical models that explicitly capture misspecifications in metabolic models
2. Develop differentiable and scalable inference algorithms using automatic differentiation
3. Implement HPC-tailored sampling strategies in Python and C++
4. Apply your framework to analyse real biological datasets to demonstrate robustness, interpretability, and practical impact
5. Contribute to open-source software tools, helping to shape future research infrastructure
6. Present your results on conferences in Germany and abroad
Your Profile:
7. Excellent Master’s degree in statistics, physics, mathematics, or a related quantitative field, ideally with a strong focus on computational practice
8. Strong mathematical and statistical background, with pronounced analytical and problem-solving skills
9. Proven programming expertise in Python and C++, with solid experience in scientific computing and software development; familiarity with Linux environments
10. Excellent collaboration and communication skills and enjoyment of working in an international, interdisciplinary research team
11. Familiarity with Bayesian thinking is desirable
12. No prior biological experience is required; curiosity for life science questions and willingness to collaborate with experimentalists is sufficient
13. Prior research experience ( internships, thesis projects, open-source