Aurelia

I am a passionate about applying statistics, machine learning and deep learning to solve real-world problems, across a variety of industries.

Our projects usually start with a proof of concept phase, and continue with pilots for testing against real IT conditions, then go into production, with an agile methodology.

Work experiences:
- Project load forecasting in the industrial sector (Python, AWS, PowerBI )
- Graph theory, community detection and influence analysis (Python Igraph, Flask, D3.JS, Javascript/Html)
- Churn B2B analysis and forecasting for a brewing company (Python, Google Cloud)
- Modelling team member during an ECB TRIM on-site inspection of Basel 2 models : Probability of Default, Exposure at Default. Models audit, statistical analyses, methodological counterarguments. (SAS, R)
- Artificial Intelligence use cases framing
- Two dashboarding projetcs (for a data center resources monitoring -Python web framework: Bottle, AngularJS, amCharts-, for an asset management company -Google Cloud storage, Dataprep, Datastudio)
- Conducted external trainings: Introduction to Machine Learning, ML algorithms with Python/with R, Data project management

Additional contributions:
- Presentations at conferences and meetups: graph theory/community detection, game theory & AI, AI in Fintechs
- Meetup "Paris Data Ladies" co-organizer: aims at promoting women in AI, in particular data scientists, engineers and architects. 10 editions so far @ Ecole 42, Google, BPCE, Deezer... +1150 members
- Involved in the recruitment process: technical & consulting interviews

Apprenticeship at the Risks Department (Valuation Division): loan performance data analysis, creation of automated quality checking tools, predictive modeling and programming.

- Design and implementation of risk prediction models for eligible european Residential Mortgage Backed Securities (prepayment and default probabilities).
- Random forest & survival analysis (Kaplan-Meier, Cox model with competing risks, random survival forest). R (caret, survival...)

- Test of a new methodology for TLTRO impact assessment, based on a paper of the Bank of Spain.
- Literature review, econometric modeling, documentation and presentation.