Samuel Rochette
Data Scientist
Pommiers, 69480, France
+33 6 15 80 54 55
Profile

Applied mathematics & computer science enthusiast with 4+ years experience. I am convinced that data scientists must have skills not only in modeling but also in clean code practices and ops to lead projects to production.

Employment History
Data Scientist at Freelance, Paris
December 2019 – Present

Mission IA for NeoXam: creation of a website that extract information from financial PDF.

Stack: python (spaCy, streamlit, scrapy, beautifulsoup4), heroku

Data Scientist at OCTO Technology, Paris
April 2016 – December 2019

Consulting:

• Faurecia - Data Science - default detection in car seat production line.

- Neural Network implementation with tensorflow to detect wrinkles zones and intensities from pictures

- Setup of AWS CI/CD

- Development of a flask-socketio application to serve the models and communicate with the camera and the prod. line protocol

- Double optimization: temporal constraint <5s and performance in recall metric

- Clean code trainer for the team

- Deployment automating with Ansible

- Lead and training of the offshore team in India

• Ministère de l'intérieur - Development - Web application to facilitate the access to french government services for foreigners

- Front-end: AngularJS / Angular5

- Back-end: Python, Flask, MongoDB

- Agile methodology (Scrum), Git Flow, clean code (Test-Driven-Development)

R&D:

Probability calibration
Deploy deep learning model on raspberry Pi

Prediction intervals in regression

• GNU Linux Mag committer (HS 100 & HS 106)

• Definition of interpretation tools for Random Forest and other black box algorithms

• Implementation of a prediction models for maintenance in industry process (Python)

Trainer at OCTO Academy:• Data Science basis training
• Data Science advanced training
• Admin Hadoop 2.X Hortonworks

Speaker:

• "Quantifying Uncertainty In Machine Learning Models", PyData New York, November 2019
• "Generating Adversarial Examples To Trick Neural Networks", DataJob Paris, November 2019

Intern at Ogury, London
May 2015 – September 2015

Statistical processing of data in order to target the ad on smartphone (Hadoop, SQL, R, java).

Education
Ecole nationale supérieure des Mines de Saint-Etienne, Saint-Etienne
September 2013 – June 2016

Degree: MSc Engineering

• 3rd year option, Big Data : Statistical classification, Optimization, Hadoop, NoSql, Data organization methods, Data mining
• Master program, Partial differential equations, Numerical scheme and simulation

• 2nd year option, Data Sciences : Statistical learning, Monte Carlo simulation, Time series, Regression analysis, Machine learning, Optimization, Kriging, Big Data

• 1st year common basis : Modeling, Probability and statistics, Computer science (C, Java, Matlab, MySQL, R), Physics, Macro/Microeconomics, Management

Association : Sports Association for students
Social project: help for reintegration into working life in jail

Technische Universität Berlin, Berlin
September 2015 – February 2016

Degree: ERASMUS exchange semester

• Engineering statistics
• Time series analysis
• Logistik
• Spieltheorie

Details
89 Montée de Corbay, Pommiers, 69480, France
+33 6 15 80 54 55
samuel.rochette06@gmail.com
Links
Skills
Python
Data Science
Vue.js
Machine Learning
Deep Learning
Docker
AWS
Languages
French
English
German
Hobbies
Skiing, Saxophone, Football, Ice hockey, Artistic Gymnastics