Jan 2020 – Jan 2021 · 1 yr · Paris, Ile-de-France, France

• In-depth analysis and implementation of Explainable AI (XAI) projects: e.g. using XAI to improve model robustness and reliance of deep learning predictions on correct features (Grad-CAM, SHAP, occlusion sensitivity); using VAE latent embeddings to measure trust in new predictions for out-of-domain samples in the context of autonomous vehicles. Some of this work was presented at the Machine Learning for Certified Systems conference, Toulouse 2021.

Jan 2015 – Jan 2019 · 4 yrs · Paris, France

• Developed a learning algorithm, which allows biological neurons to represent visual information efficiently. The unsupervised learning rule incorporates specific biological constraints: realistic synaptic connection probabilities in local cortical networks and synaptic delays.
• Presented my research regularly at international conferences in France, Portugal, Germany and U.S.A.
• Was one of the lecturers for the Cognitive Science Masters course: “Machine Learning Applied to Neuroscience”, at École Normale Supérieure, for the years 2018 and 2019.

Jan 2011 – Jan 2015 · 4 yrs · Vancouver, Canada

• Member of the Advanced Analytics team on the Smart Metering and Infrastructure project
• Worked on developing new algorithms to detect anomalies, such as energy theft and network connectivity errors from the large quantity of power consumption data recorded by smart meters.
• The algorithms I helped develop are new in the electric utility field and use discrete measurements of voltage and power at leaf nodes, along with the impedance of electrically conducting network edges, to compute upstream voltage series and perform cluster analysis.
• I worked on identification of business needs and benefits, developed mathematical solutions, wrote prototype programs, and trained end users after the release of our applications. I coauthored IEEE peer reviewed papers, and regularly made presentations for technical and non-technical audiences.

Oct 2009 – Dec 2010 · 1 yr 2 mos · Vancouver, Canada

• Worked in a renewable energy research team on the development and validation of the tidal turbine computational fluid dynamics model.
• Successfully applied for grants, supervised engineering interns, and liaised between the company and our university and industry collaborators.

Jul 2008 – Jul 2008 · 0 mo · Vancouver, Canada

• Led a team of five international undergraduate students during an industrial project proposed by Pulse Energy, a company based in Vancouver.
• My team developed an algorithm for determining typical energy use in buildings, so that atypical use could be detected and corrected, thus leading to a more sustainable energy use.

May 2007 – Aug 2007 · 3 mos · Montreal, Canada

• Worked with Dr. Jéan Côte from Environment Canada on improving the dynamic core of the Canadian regional climate model.
• Helped implement a new scheme for the solution of the shallow water equations on a rotating sphere. The scheme uses a Yin-Yang grid (two overlapping rectangles) to cover the sphere, rather than the usual longitude-latitude grid, thus removing the apparent singularity at the poles.

Jan 2006 – Aug 2006 · 7 mos · Vancouver, Canada

• Worked with Dr. Mary Catherine Kropinski on finding numerical schemes which best approximate solutions of a 4th order nonlinear partial differential equation. This equation models the microphase separation of diblock copolymers.

Jan 2005 – Dec 2005 · 11 mos · Vancouver, Canada

• Created an interactive website which performs data analysis on the SQL server and returns electricity demand trend graphics to the user who monitors these trends.
• Performance testing of a software for electricity demand forecast and determining the accuracy of distribution line current readings
• ­- Helped develop an in-house SAS based software which estimates future electricity demand trends for different types of customers based on statistical analysis