About
I'm a researcher working at the intersection of statistics, machine learning, and drug discovery and development. My work spans methodology and application: causal inference, design of experiments, and AI/ML methods on one side, and multi-omics foundation models, clinical biomarker discovery, target identification and assessment, and agentic AI for drug discovery and development on the other. I completed my PhD in Statistics at EPFL, and have since spent five years bringing rigorous statistical and computational methods to problems across the drug development pipeline.
Publications
-
Cross-modal transfer learning for mapping bulk transcriptomes at cellular level
3rd Workshop on Multi-modal Foundation Models and Large Language Models
-
TSAssistant: A Human-in-the-Loop Agentic Framework for Automated Target Safety Assessment
*Equal contribution and corresponding authors
AI for Science (AI4Science) Workshop
-
Quantitative clinical pharmacology supports totality of evidence, data integration and dose selection proposal of a new molecular entity with no single agent activity
-
Distinct mechanisms of CD28 and 4-1BB costimulation in glofitamab combination therapies for relapsed/refractory non-Hodgkin lymphoma (R/R NHL)
-
CD19-CD28 (RO7443904) combination with glofitamab enhances T-cell proliferation and effector function in patients with relapsed/refractory non-Hodgkin lymphoma (R/R NHL)
-
Clinical relevance of spatially-resolved transcriptional signatures characterizing macrophages from the light and dark zone of the germinal centre
-
Multiple testing using uniform filtering of ordered p-values
-
Implications of ocular confounding factors for aqueous humor proteomic and metabolomic analyses in retinal diseases
-
RO7119929, a TLR7 agonist prodrug, induces local inflammation of the tumor microenvironment (TME) by reprogramming myeloid cells in patients with advanced primary or metastatic liver cancer
-
Englumafusp alfa (CD19-4-1BBL) and glofitamab combination in patients with relapsed/refractory non-Hodgkin lymphoma (R/R NHL): biomarker results from a phase I dose-escalation study
-
Correlation of aqueous, vitreous, and serum protein levels in patients with retinal diseases
-
Deciphering molecular and cellular ex vivo responses to bispecific antibodies PD1-TIM3 and PD1-LAG3 in human tumors
-
Multiple testing with test statistics following heavy-tailed distributions
Experience
-
2021 — Present
Senior Scientist
F. Hoffmann-La Roche · Basel, Switzerland
-
2016 — 2021
Ph.D. in Statistics
École Polytechnique Fédérale de Lausanne (EPFL) · Lausanne, Switzerland
-
2014 — 2016
M.Sc. in Statistics
University of Science and Technology of China (USTC) · Hefei, China
-
2010 — 2014
B.Sc. in Statistics
University of Science and Technology of China (USTC) · Hefei, China
Featured Work
TSAssistant
A human-in-the-loop agentic framework that automates target safety assessment for drug discovery, pairing LLM-based agents with expert oversight.
Agentic AICross-modal Transcriptome Mapping
A cross-modal transfer learning approach that maps bulk transcriptomes to cellular-level resolution using multi-modal foundation models.
Foundation models