I am a senior researcher specialising in Bayesian Statistics at the Applied Maths and Computer Science Department of INRAE, France.

I graduated from Ecole Normale Supérieure de Cachan, near Paris, and from Ecole Polytechnique Fédérale de Lausanne, both in physics. During my master (2009-2010), I participated in an exchange program with the National Taiwan University where I worked in the group of Jer-Lai Kuo, at the Institute for Atomic and Molecular Science. While in Lausanne (2010-2012), I worked in the lab of Giuseppe Foffi in collaboration with Davide Fiocco and Hans-Joerg Limbach, and in the lab of Félix Naef in collaboration with Jonathan Bieler. In April 2012, I moved to the Institute for Scientific Interchange, in Turin, for an internship in the group of Vittoria Colizza, where I worked with Paolo Bajardi and Michele Tizzoni. In October 2012, I joined the Laboratoire de Biométrie et de Biologie Evolutive at Université Claude Bernard, Lyon to start a Ph.D. under the supervision of Sandrine Charles and Marie-Laure Delignette-Muller, which I defended in October 2015. During this Ph.D. in biostatistics applied to ecotoxicology, I studied the ecological risk of contaminants (e.g.pesticides). In collaboration with experimental biologists, I developed several improvements to the analysis of ecotoxicological data. I later oriented on more theoretical topics as a post-doctorate. I joined the Bayesian Nonparametrics group of Igor Prünster and Matteo Ruggiero, at the Collegio Carlo Alberto and the Università degli Studi di Torino. Since January 2020, I am a permanent researcher at the Applied Maths and Computer Science Department of INRAE.

My scientific contributions concern censored data analysis, Bayesian hierarchical modelling, survival data analysis, Bayesian Nonparametric (BNP) density estimation and mixture models, dependent BNP processes and inference for hidden diffusion processes.

Overview of PhD. research

 My research focuses on tackling biological problems using quantitative methods from statistics and physics. I did my Ph.D. in biostatistics applied to ecotoxicology, the study of the ecological risk of contaminants (e.g. pesticides). In collaboration with experimental biologists, I developed several improvements to the analysis of ecotoxicological data. This data is expensive to collect, yet a large proportion is usually discarded or entirely summarized, which is both wasteful and statistically questionable. My work was divided in three articles:

1) I showed the deleterious influence of discarding censored ecotoxicological data (i.e. resulting from experiments which did not produce a significant effect) and presented a method to take them into account. I contributed to a web-tool allowing the scientific community to use this methodology and perform a statistically sound ecological risk assessment.

2) The traditional approach of summarising the raw data discards valuable biological information and underestimates uncertainty and variability. I developed a bayesian hierarchical model that includes all the information available in the raw data and constructed a new quantitative indicator of ecological risk.

3) Usually, only data at the end of the experiment is considered, resulting in a risk assessment dependent on the duration of experiments. I showed how to take time-resolved data into account and how to derive a time-resolved risk indicator.

I am currently developing this model to predict this risk indicator for arbitrary time-resolved contamination scenarios. I am also involved in a project concerning the extrapolation of these approaches to longer-term impacts of contamination.



Currently supervising:

Past supervisions:

Funded projects and grants

Fellowships and awards