Nicolas Guigui, PhD

Founded as a spin-off from CNRS lab Institut Langevin, OWLO is building next generation microscopy for healthcare. Our unique technology is the first to provide real-time 3D imaging in a non invasive way. This innovation will change research and medical practice by making visible and quantifying what was previously unseen. Our first use case is in Assisted Reproductive Technologies (ART) where 75% of embryos transfers fail today, resulting in long and expensive procedures that are very stressful and demanding for women. We provide an entirely new way of looking at embryos with new indicators to predict their viability before implantation to improve the efficiency of the procedure and reduce its cost.

Implementing real time 3d imaging algorithms in CUDA/C++

I am a member of the Medical Imaging team, Epione, and of the ERC project G-statistics. As such, I am working on computational tools for statistics on manifolds, and applications to shape analysis. I am also involved in the development of geomstats, a python package for geometry and machine learning --> More info on github or at geomstats.ai For a list of publications see my scholar profile. All full texts are available on hal.

Designing methods to integrate gene-gene interaction graphs in models of imaging-genetic data to study Neurological disorders. This work was presented at ISBI 2019 in Venice, see presentation below and online paper at https://hal-cea.archives-ouvertes.fr/cea-02016625

As part of the R&D Platform Technology and Science entity, my position involves two main projects. Automation of tissue-section segmentation from optical images: delivered an end-to-end solution to automatically segment region of interest on histopathology slices: • Gathered, curated and labelled a training dataset • Automated the extraction of tissue from background (with OpenCV) • Trained an object detection deep learning architecture to remove specific regions of tissues (in Keras) • Deployed the model for daily use Exploratory Data Analysis for computational toxicology: • Used Python and relevant data analysis packages (Pandas, Numpy, Scikit-Learn, Keras) • Learnt to convince non-mathematicians with data-driven results for decision making • Explored complex datasets with unsupervised learning methods, combining data from multiple sources and types • Delivered proof-of-concept and internal tutorials

Statistical Theory - Applied Statistics - Modern Statistical Methods - Gaussian Processes - Convex Optimisation - Advanced Probabilities - Stochastic Calculus - Biostatistics. Dissertation: "Exploring the Statistical utility of Deep Learning"

Measure Theory - Statistics - PDE - Algorithmics and Programation in Python - Heat Transfer - Nuclear Energy- Economy - Financial Risk Modelling - Corporate Finance - Information Technology - Biology - Epistemology - Quantum Physics -Mechanics - Thermodynamics - Project Management - Business Management Théorie de la mesure - Statistiques - Equations aux Derivées Partielles - Algorithmique et Programmation - Transferts Thermiques - Neutronique - Economie - Modélisation des Risques Financiers - Finance d'entreprise - Sciences de l'information - Biologie - Epistémologie - Physique Quantique - Mécanique - Thermodynamique - Gestion de Projets - Ingénierie des Systèmes Complexes - Gestion d'entreprise