Paul Lemaitre Email & Phone Number

Montreal, Quebec, Canada

I am currently a Senior Data Scientist at IVADO Labs, a world-class AI consulting company based in Montreal. I have been working on short to mid-length (6-12 months) projects that are related to various industries such as:- supply chain and spare part forecasting for aeronautics- medical imaging- precision agriculture- surgery scheduling for hospital bedsMy core strengths include Machine Learning, Computer Vision and Uncertainty Quantification. I am a… Show more I am currently a Senior Data Scientist at IVADO Labs, a world-class AI consulting company based in Montreal. I have been working on short to mid-length (6-12 months) projects that are related to various industries such as:- supply chain and spare part forecasting for aeronautics- medical imaging- precision agriculture- surgery scheduling for hospital bedsMy core strengths include Machine Learning, Computer Vision and Uncertainty Quantification. I am a seasoned Python developer and am proficient in Data Science related libraries such as numpy, pandas and Tensorflow. Show less

Montreal, Canada Area

As a Research Associate in a medical neuroimaging research laboratory (10+ students), I helped scaling up the lab with setting up the deep learning environment (NVIDIA DGX / GPUs) necessary to pursue the research. Throughout my time at PVG, I assisted students in their researches on convolutional neural networks via code reviews, mentoring, and research conducting. The core of the lab's research was about 3d multimodal segmentation of MS lesions / brain tumors.I also assisted… Show more As a Research Associate in a medical neuroimaging research laboratory (10+ students), I helped scaling up the lab with setting up the deep learning environment (NVIDIA DGX / GPUs) necessary to pursue the research. Throughout my time at PVG, I assisted students in their researches on convolutional neural networks via code reviews, mentoring, and research conducting. The core of the lab's research was about 3d multimodal segmentation of MS lesions / brain tumors.I also assisted the professor in various lab tasks (grants, reports, presentations). Show less

Région De Montréal, Canada

As a Data Scientist in a leading chemistry company based in Montreal, I: - Applied anomaly detection for multi-dimensional time series (Analysis and modelling of chlorine factories start/shutdown) - Developed as a team an open source full research framework for asynchrone experiment planning for multiple researchers and multiple GPU servers. - Implemented research papers on Gaussian processes for automatic hyperparameters tuning and state of the art Deep Learning research… Show more As a Data Scientist in a leading chemistry company based in Montreal, I: - Applied anomaly detection for multi-dimensional time series (Analysis and modelling of chlorine factories start/shutdown) - Developed as a team an open source full research framework for asynchrone experiment planning for multiple researchers and multiple GPU servers. - Implemented research papers on Gaussian processes for automatic hyperparameters tuning and state of the art Deep Learning research (Variational Dropout in keras). I worked mostly in Python, with the classical packages for data science (numpy, pandas, seaborn, keras, scikit-learn). Other tools included docker (nvidia-docker) and git. Show less

Toronto/Montreal

At this position in the Data Science team of a freemium video game company, I focused on predicting Key Performance Indicators (KPIs) for game adoption and monetization.To do so, I:- Used statistical learning techniques such as Deep and Recurrent Neural Networks, General Linear Models, and Survival Analysis .- Produced data visualization for leadership- Tuned hyperparameters using metamodel-based approaches such as kriging and the Efficient Global Optimization (EGO)… Show more At this position in the Data Science team of a freemium video game company, I focused on predicting Key Performance Indicators (KPIs) for game adoption and monetization.To do so, I:- Used statistical learning techniques such as Deep and Recurrent Neural Networks, General Linear Models, and Survival Analysis .- Produced data visualization for leadership- Tuned hyperparameters using metamodel-based approaches such as kriging and the Efficient Global Optimization (EGO) algorithm.- Applied collaborative filtering for recommending tools to non-specialized users.In the back-end, I got proficient in managing and querying MongoDB databases, alongside with acquiring familiarity with an in-house Hadoop-like querying systemThe stack used was mostly Python, using libraries such as numpy, sklearn, theano, keras, mongoengine, flask, and pandas. Some of the research work was conducted in R, specifically using packages such as survival and forecast. Show less

Lyon, France

I was involved in a complex system modelling platform development and in Big Data management.In the first context, I worked in a team in which we develop and implement optimization and sensitivity analysis algorithms in R and on a Linux and Windows environment. We also were interested in surrogate functions for time-consuming models. In the second context, we manipulated large data set and feed the modelling environment with the correlations extracted. We work in R and on the… Show more I was involved in a complex system modelling platform development and in Big Data management.In the first context, I worked in a team in which we develop and implement optimization and sensitivity analysis algorithms in R and on a Linux and Windows environment. We also were interested in surrogate functions for time-consuming models. In the second context, we manipulated large data set and feed the modelling environment with the correlations extracted. We work in R and on the Hadoop framework. Show less

Chatou, France

This applied PhD thesis primarily focuses on sensitivity analysis (SA) when the quantity of interest is a binary function (system failure or success). The objective of SA is to identify the contribution of input uncertainty to output uncertainty. Most techniques are designed for continuous quantities of interest. The available methods for probabilistic quantities of interest often rely on restrictive assumptions or require a large number of simulation function calls, which can be… Show more This applied PhD thesis primarily focuses on sensitivity analysis (SA) when the quantity of interest is a binary function (system failure or success). The objective of SA is to identify the contribution of input uncertainty to output uncertainty. Most techniques are designed for continuous quantities of interest. The available methods for probabilistic quantities of interest often rely on restrictive assumptions or require a large number of simulation function calls, which can be computationally expensive. Thus, the aim of this work is to develop adapted methods to assess the impact of uncertainty on multiple inputs for a binary output. Throughout this research, we tested existing sensitivity analysis methods and proposed more efficient original approaches.The highlights of this thesis include:- A comprehensive state-of-the-art review on sensitivity analysis (with over 1000 citations)- A novel and general methodology (DMBRSI/PLI) that reflects the impact of input density modification on the probability of failure - Application of the developed methods to a real case in nuclear safetyThe entire work was conducted using R (with a contribution to the development of the sensitivity package) and OpenTURNS, a Python module for uncertainty treatment. Show less

Use of wavelet shrinkage and local polynoms interpolation to smooth noisy images of mechanical deformation.

Cadarache, France

Implementation of sensitivity analysis methods in R.

Québec, Canada

Implementation of POD methods in C++.

Doctorat (Phd), Applied Mathematics, With Highest Honors

In the context of studying a numerical model representing a complex physical system, it is essential to focus on quantifying the impact.

Master Of Engineering, Applied Mathematics / Statistics

Master Of Research, Probability / Statistics

M.Sc., Statistics