Freelancer profile translated to English.

Description

Data science and artificial intelligence engineer for over 6 years, I have worked on various projects and in different sectors, particularly in image, text, video processing, and voice recognition. In the missions I have carried out, I manage the data processing process (structured or unstructured) from the preprocessing phase to the deployment of the predictive model.

Languages

French
Native or bilingual

Workplace preferences

Can work on-site

Paris (up to 50km)

ALTEN
Tech Lead Data Scientist
DIGITAL AND IT
February 2022 - Today (4 years and 4 months)
Sèvres, France
Tech Lead Data Scientist

Analysis of presidential debates
• Identify speakers in presidential debates using GMM and i-vectors.
• For each speaker, calculate speaking time.
• For each speaker, retrieve text and apply automatic text summarization models.
• Identify thematic changes in discourse using clustering approaches.
• Evaluate each part using metrics.
• Deploy results via a Flask API in Azure using Docker.
Fake news detection
• Identify fake news to ensure the integrity of information on the internet.
• Develop neural networks to detect fake news in textual data.
• Extract features such as heartbeats, lip movements to identify fake news in videos.
• Create a meta-model that allows identifying fake news in text and video simultaneously.
Scoring model
• Identify defaulting customers using a scoring model.
• Data exploration and cleaning.
• Feature engineering and class balancing.
• Develop several classification approaches.
• Model evaluation and optimization.
• Model interpretation.
• Develop a dashboard with Flask and Streamlit.
• Deploy the Dashboard in Azure via a CI/CD pipeline.
Python Azure Docker DevOps Flask
ATLEN
Predictive Maintenance with Artificial Intelligence
DIGITAL AND IT
March 2021 - January 2022 (11 months)
Sèvres, France
Analysis and modeling of vibration data to learn vibration behavior in order to anticipate anomalies for predictive maintenance and avoid unnecessary maintenance.
• Analysis and modeling of vibration data to learn vibration behavior in order to anticipate anomalies for predictive maintenance and avoid unnecessary maintenance.
• Retrieve vibration data from Raspberry via the I2C bus.
• Exploratory data analysis and cleaning.
• Feature extraction describing the signal using Fourier transform.
• Test several modeling approaches to detect anomalies.
• Model evaluation, optimization, and interpretation.
• Create dashboard with Streamlit and deploy in Azure.
• Create the database via MySQL.
• Manage daily meetings and sprint planning.
• Manage Kanban.
ALTEN
Tech Lead Data Scientist
DIGITAL AND IT
April 2019 - February 2021 (1 year and 11 months)
Sèvres, France
Improving autonomous driving:
• Analyze traffic videos to predict the intention of pedestrians to cross or not at the crossing point.
• Extract image sequences for each pedestrian from videos.
• Extract features such as pedestrian postures and others.
• Train models on bounding boxes, postures, and VGG features.
• Create a meta-model to predict pedestrian intention using the three features simultaneously.
• Deploy in Azure via Flask with Docker.

Behavior analysis in videos:
• Improve vehicle cockpits through user experience.
• Analyze videos of interaction between users and their cockpits to predict different emotions expressed by users using Deep Learning and Machine Learning.
• Identify emotion through textual, visual, and audio data.
• Extract features for audio, clean and vectorize text, extract features for images.
• Create base models for each modality.
• Create a meta-model to predict emotion across the three modalities.
• Deploy in Azure via CI/CD pipelines.
Deep Learning Deployment Python PySpark OpenCV Azure DevOps