cv

PhD - Online Continuous Image Classification with Memory-based Methods, application to YouTube data

• Online Continual Learning for Image Classification
• Memory-based methods
• Probabilistic Modeling of the Cross-Entropy loss function
• Stability-plasticity trade-off
• Collection and experiments on real YouTube data

Exchange Semester

• Exchange semester at the University of Tokyo, 1st Japanese university, in the Information Science and Technology department.
• Integration of the AIZAWA-YAMASAKI lab.
• First author of a domestic research paper and speaker at an IEICE conference in Okinawa, Japan.
• Computer Vision, NLP, Machine Learning and Differential Privacy.

Engineering Degree

• Major in Computer Science for Decision Support
• Optimization Algorithms (traveling salesman problem, etc.)
• Statistics and Probabilities
• Project Management

Classe Préparatoire MPSI/MP*

• Intensive Maths and Physics studies

Teaching Assistant

• 3rd year classes (L3)
  • Data science (python, numpy, sklearn)
  • Optimisation et IA (python, keras)
  • Introduction to Deep Learning (python, keras)
  • Algorithms and Data structures (C)
  • Project supervisor:FootTracker
• 1st year classes (L1)
  • Micro-processors (ARM assembly language)
  • Introduction to Object-Oriented Programming (java)

Visiting Scientist

• Visiting Scientist in the Computer Vision and Media Lab (CVM)
• Worked with Prof. Yamasaki in the YAMASAKI-lab
• Research topics
  • Online Continual Learning
  • Mutual Learning
  • Adversarial Training
  • Impact of Image Generation on Continual Learners

Data Science Mentoring

• OpenClassrooms \& CentralSupélec (M2)
• Management of data science projects and scientific follow-up
• Projects include data processing, univariate analysis, multivariate analysis, data visualization, supervised ML, RF, SVM, XGB, data encoding, unsupervised ML, t-SNE, k-means, DBSCAN, hierarchical clustering, transfer learning, SIFT, ORB, LSTM, cloud model deployment (AWS)

Data Scientist

• Digital Services Factory (DSF) of Faurecia, one of the first global automotive suppliers worldwide
• The DSF team, composed of 60 engineers aims to bring new AI skills.
• pecific involvements include ML applied to verbal aggressivity detection, DL applied to driver distraction detection and passenger presence detection. Model optimization. Inference on NVIDIA Jetson boards.

Research Engineer

• Part of the innovation team of Wassa, a digital company of 30 employees.
• The R&D team, composed of 4 researchers, ensures the quality of Wassa algorithms by using state-of-the-art CV techniques.
• Specific involvements include Deep Learning applied to Gender Recognition, Age Prediction and Object Detection. HoG for pedestrian detection, TS forecasting.

Deep Learning Intern

• Part-time research internship at So-net Media Networks (SMN), a SONY subsidiary of 1000 employees specialized in Online Advertising.
• Implementation of a tool in Python to classify advertisement images regarding their efficiency.
• Training of several Deep Learning models using PyTorch.
• Experience training on a remote GPU using CUDA.
• Reading and understanding several Deep Learning research papers.
• Worked with a database of 40,000 ad images.

Data science

• Python, Numpy, Sklearn. Original exercices from Jean-François Bercher

Optimisation et IA

• Python, Keras

Algorithms and Data structures

• C programming, complexity, code coverage

Introduction to Deep Learning

• Python, Keras, face detection and recognition

Project Supervisor of FootTracker

• Analyzing soccer (football in French) to extract interesting stastics given player positions.

Object-Oriented Programming

• Introduciton to Java OOP

Micro-processing ARM

• ARM assembly programming

Project Supervisor of Youtube DeepMetaData

Project Supervisor of YouTube Catalogue.

• Supervising student working on scraping and organizing YouTube data.

Time Series & Machine Learning

• Teaching Machine Learning concepts to final year engineer students for time series forecasting (LSTM, transformers, etc.)