Simon Bozonnet

- Lead technically the AI-Factory on Natural Language Processing and Computer Vision (Document AI). It includes technical reviews, feasibility study and cost estimates, - Setup the technical roadmap, the ML strategy to expose AI cutting-edge technological services and the ML environments (ML Data Scientist sandbox, ML Platform for production) - Supervise Data Science activities: define scientific and methodological strategy, recruit and manage Data Scientist team, provide a technological watch of latest research progress - Design a Data Scientist work environment on AWS and a data annotation tool - Set up and lead a data annotation unit to address AI requirements - Evangelize AI toward business team

Design and Implementation of the following systems:

- Automatic Email Classification: machine learning based system for automatic email classification.

- Handwriting Recognition/OCR: deep learning system (sequence to sequence LSTM based) for handwriting and typed characters on specific insurance documents

- Computer Vision: automatic document classification (supervised/unsupervised), pattern detection and signature detection through a deep learning system.

- Fraud Detection in health insurance: unsupervised detection of outliers and customer behavior modelling.

- Fraud Detection at PayPal: feature analysis and data modeling to detect suspicious activities

- Text classification at Bearingpoint Gmbh: text classification for file merging

- AstraZeneca challenge: predict the best medicine for a specific patient.

- Unsupervised customer clustering from customer survey: predict the success and the segment of some start-ups from their customer survey.