About About | Code · Data Science

  • Currently working on a book: "Essential Math for Data Science"
  • Previously working as a Machine Learning Scientist at Ava Accessibility ava icon.
  • Previously Ph.D student at the École Normale Supérieure ens icon

Data Science

python icon Python
r icon R
Matplotlib icon Matplotlib
D3 icon D3
GGPlot2 icon GGPlot2
sklearn icon Scikit Learn
keras icon Keras
Pandas icon Pandas
numpy icon Numpy
Scipy icon Scipy

I am currently managing a project for bird detection using deep learning with the non profit organization Wazo in Paris. This project has been selected in the season 06 of DataForGood Paris from September to December 2019.

I am also bloging here on mathematics for machine learning and deep learning. I think that computer science is a great way to learn theoretical knowledge with a practical approach.

At Ava, I worked on creating and maintaining machine learning pipelines for speaker diarization from multi-microphone signals.

I used R to analyse behavioral data and create vizualisations and Python to analyse EEG data (see my toolbox for EEG processing) and elaborate offline/online signal processing workflow.

At the corner of data science and web developement, I created the skeleton of a neurofeedback app that streamed and transfered the data from the EEG system to a web server in Django and get the data in the browser with web sockets for final feedback display.

Web Development - Full-Stack

django icon Django
flask icon Flask
digitalocean icon Digitalocean
heroku icon Heroku
javascript icon Javascript
html icon HTML
css icon CSS
react icon React
couchdb icon CouchDB
pouchdb icon Pouchdb

During my PhD, I have developped Web Apps using Django and Javascript for auditory experiments running on computers and tablets. I also worked with NoSQL databases (CouchDB) hosted on a DigitalOcean and PouchDB to build offline-first web app.

I am also using D3 and React to build data vizualisation on the web.

I have created Web Apps using the Web Audio API to create sounds with controled acousticx features (for instance a demo of amplitude and frequency modulations with visualizations).

Some Projects

Mechanism of the gradient descent algorithm
Linear Algebra Series
Series on linear algebra chapters from the Deep Learning Book from Ian Goodfellow, Yoshua Bengio, and Aaron Courville (2016): 12 blog posts + Python Notebooks to get all you need to do great things in machine learning and deep learning.
Screenshot of amplitude and frequency modulation demo
Amplitude and Frequency modulations
Interactive demo illustrating the concept of amplitude and frequency modulation.
Screenshot of tone cloud demo
Tone clouds
Audiovisual represention of tone clouds. Tone clouds are sets of pure tones with random spectrotemporal characteristics.