Freelancer profile translated to English.

Description

Computer engineer specialized in Machine Learning. I can join a team or take charge of an AI project from A to Z, namely:

Data retrieval:

- Collect or create data from your sources (crawlers, scraping, internal/external APIs, databases, spreadsheets, ...)

Data processing:

- Read files in the following formats: json, csv, text, excel

- Clean and process data with Pandas and Numpy

- Store in a database as needed, I use the most appropriate DB for the project (mongoDB, PostgreSQL, Neo4J)

Data visualization:

- Visualization with Python: matplotlib, seaborn, plotly

- Or in dynamic dashboards: Dash/Bokeh

Data analysis, prediction, and classification:

- Preliminary data analysis with sklearn

- If necessary, train a Deep Learning model with Tensorflow, Pytorch, or Keras depending on the application

Creation of the final service/deliverable:

- Create APIs with Flask, FastAPI

- Create applications with Flask, deploy on AWS or self-hosted

Languages

French
Native or bilingual
English
Native or bilingual

Workplace preferences

Remote only

Primarily works remotely

NVIDIA
Machine Learning Intern
January 2020 - July 2020 (6 months)
Austin, TX, USA
- Contributor to the RAPIDS cuML machine learning open source repository. RAPIDS is an open-source platform incubated at NVIDIA that enables end-to end data science workflows on GPUs. cuML's API is based on popular projects such as Scikit-Learn and Dask.
- Wrote a blog post for NVIDIA about my work on accelerating NLP preprocessing on GPU: https://medium.com/rapids-ai/natural-language-processing-text-preprocessing-and-vectorizing-at-rocking-speed-with-rapids-cuml-74b8d751812e
Criteo
Deep Learning Engineer Intern
February 2019 - August 2019 (6 months)
Paris, France
Implemented self-attention mechanism to recommender systems with a variation of The Transformer model from the paper Attention is all you Need. Conducted offline evaluation of the model with significant metrics uplift (+12%) on prospecting scenarios. Written in Python and Tensorflow. Large dataset of about 300 millions samples.
SOLUTECH SAS
Software Engineering Intern
August 2017 - February 2018 (6 months)
St Barthélemy
- Designed and implemented algorithms for optic fiber deployment with Python and GIS databases with the goal to win an invitation to tender for the market of optic fiber deployment in St Barthélemy. Automatic generation of the graph describing the infrastructure of cables throughout the island. Optimization of the graph given interchangeable objective criterions through exhaustive search and pruning.
- Development of an android app for geolocalized data collection with dynamic forms generated from config files served by a backend server. This app was used by architects to facilitate collection of critical data such as water pipelines and electrical lines status after the hurricane Irma struck the islands of St Barthélemy and St Martin.