Aymeric Roucher
About Aymeric Roucher
Aymeric Roucher is a Product Data Scientist at Descartes Underwriting, where he develops new climate insurance products. He has a background in data science and machine learning, with experience at various organizations including IAC Partners and the University of Cambridge.
Work at Descartes Underwriting
Aymeric Roucher has been employed at Descartes Underwriting since 2021 as a Product Data Scientist. In this role, he focuses on developing new climate insurance products. He has collaborated with one of the cofounders of Descartes Underwriting to conduct market studies aimed at product design. Additionally, he has spearheaded the industrialization and global generalization of the company's in-house cyclone model.
Education and Expertise
Aymeric Roucher has a strong academic background in mathematics and computer science. He studied at Lycée Henri IV, where he achieved his Baccalauréat from 2012 to 2015. He then attended Lycée Sainte-Geneviève, earning a BSc in Mathematics & Physics from 2015 to 2017. Roucher furthered his education at École Polytechnique, obtaining an MSc in Computer Science from 2017 to 2021. He also completed an MPhil in ISMM/Machine Learning & Machine Intelligence at the University of Cambridge from 2020 to 2021.
Professional Experience
Before joining Descartes Underwriting, Aymeric Roucher gained diverse experience in various roles. He worked as a Consulting Intern at IAC Partners in 2019 for two months in Chicago. He served as a Data Science Consultant at the University of Cambridge from 2020 to 2021 for three months. Additionally, he held positions as a Machine Learning Engineer Intern at Wandercraft in 2020 for four months and as a Visiting Data Science Researcher at Chateau Latour in 2021 for eleven months. He also served as a Lieutenant in the Armée de Terre from 2017 to 2018.
Achievements
Aymeric Roucher has made significant contributions in the field of artificial intelligence and machine learning. He won a €5,000 prize in an AI challenge by successfully teaching a swarm of drones to autonomously navigate their environment. He also developed a machine learning model that became standard practice for estimating trajectories from incomplete sensor data, completing the project in five months, significantly ahead of the initial one-year estimate.