The goal of team VPE is to develop a versatile computational platform that can predict the efficacy of first- or best-in-class drug candidates in virtual patient populations at an unprecedented accuracy, thereby addressing one of the most critical bottlenecks of the pharmaceutical industry today: a 90% failure rate of new drug candidates during clinical development. In partnership with Sanofi, the VPE team will develop innovative artificial intelligence methods to build the virtual patient platform. As a proof-of-concept, the initial platform will focus on chronic immune-mediated diseases such as atopic dermatitis (AD) and inflammatory bowel disease (IBD), where new medication that can address patient heterogeneity is needed.
Mentors
- Dr. Tommaso Andreani (Industry Mentor)
Sr. Principal Data Scientist – Disease Modeling Innovation Lead, Sanofi
- Dr. Jan Korbel (Academic Mentor)
Head of Data Science, European Molecular Biology Laboratory
Publications
- Ansh Kumar, Sanjana Balaji Kuttae, Charlie George Barker, Thomas Rückle, Anastasios Siokis, Sven Jager, Thomas Klabunde, Tommaso Andreani, Gurdeep Singh, and Ahmad Wisnu Mulyadi (2026). State-aware policy optimization for a reliable multi-turn, multi-tool scientific agent in kinetic biological models. International Conference on Machine Learning (ICML) Workshop: AI4Science 2026
Read paper
- Charles George Barker*, Ahmad Wisnu Mulyadi*, Sanjana Balaji Kuttae, Ansh Kumar, Lilija Wehling, Thomas Rückle, Sommer Anjum, Anastasios Siokis, Hans-Christoph Schneider, Sven Jager, Thomas Klabunde, Tommaso Andreani, and Gurdeep Singh (2026). Regimen-aware forecasting for mechanistic virtual patients with time-series foundation models. ICLR 2026 Workshop on Time Series in the Age of Large Models (TSALM)
*equal contributions
Read paper
- Ahmad Wisnu Mulyadi*, Charlie George Barker*, Sanjana Balaji Kuttae, Lilija Wehling, Thomas Rückle, Nicolas Boucher, Firas Abdessalem, Sven Jager, Anastasios Siokis, Sommer Anjum, Mohammed H. Mosa, Thomas Klabunde, Tommaso Andreani, and Gurdeep Singh (2025). Evaluating time-series foundation models as zero-shot surrogates for mechanistic virtual patients. European Conference on Neural Information Processing Systems (EurIPS) Workshop: SimBioChem 2025
*equal contributions
Read paper
- Lilija Wehling*, Gurdeep Singh*, Ahmad Wisnu Mulyadi, Rakesh Hadne Sreenath, Henning Hermjakob, Tung V. N. Nguyen, Thomas Rückle, Mohammed H. Mosa, Henrik Cordes, Tommaso Andreani, Thomas Klabunde, Rahuman S. Malik Sheriff, and Douglas McCloskey (2025). Talk2Biomodels: AI agent-based open-source LLM initiative for kinetic biological models. BMC Bioinformatics
*equal contributions
Read paper
- Ahmad Wisnu Mulyadi, Lilija Wehling, Ansh Kumar, Nicolas Boucher, Firas Abdessalem, Sven Jager, Mohammed H. Mosa, Thomas Klabunde, Tommaso Andreani, and Gurdeep Singh (2025). BioMedReasoner: Towards multi-hop reasoning using path-based relational learning on biomedical knowledge graphs. AI4Science Workshop at NeurIPS (2025)
Read paper
- Gurdeep Singh*, Lilija Wehling*, Ahmad Wisnu Mulyadi*, Rakesh Hadne Sreenath, Thomas Klabunde, Tommaso Andreani, and Douglas McCloskey (2025). Talk2Biomodels and Talk2KnowledgeGraphs: AI agent-based application for prediction of patient biomarkers and reasoning over biomedical knowledge graphs. MLGenX 2025, ICLR Conference Workshop Paper
*equal contributions
Read paper
You can also find us on GitHub
The research of this team is kindly supported by Sanofi.
