Fabien Gadet

Data scientist / engineer

Moves to Paris

  • 48.8546
  • 2.3477
Propose a project The project will only begin when you accept Fabien's quote.
Propose a project The project will only begin when you accept Fabien's quote.

Location and geographical scope

Location
Paris, France
Can work in your office at
  • Paris and 10km around

Preferences

Project length
  • ≤ 1 week
  • ≤ 1 month
  • Between 1-3 months
  • Between 3-6 months
  • ≥ 6 months
Company size
  • 1 person
  • 2 - 10 people
  • 11 - 49 people
  • 50 - 249 people
  • 250 - 999 people
+2 autres

Verifications

Influence

Languages

Categories

Skills (35)

Fabien in a few words

Data scientist / engineer depuis 2 ans (MASTER software engineering), j'ai réalisé de nombreux projets autour de la data autant du coté database, ETL pipeline que du coté analyse et data science .

J'ai travaillé pendant 1 an et demi chez Qobuz en tant que data scientist , j'ai réalisé de a à z plusieurs modèles prédictifs afin d'aider à la rétention, conversion des clients et détection de fraude et pour alimenter une data warehouse complète. (+5 à 10% de rétention supplémentaire suite à la mise en place de mes modèles)

J'ai également de l'expérience en web (API, architecture) et plusieurs année de développement en C , C++

Réalisation de plusieurs projets autour de la data sur mon temps libre (principalement en Julia / Python)

Experience

Qobuz

High Tech

Data Scientist

Région de Paris, France

September 2019 - Today

• Data Engineer/ analysis:
o ETL pipelines to follow churn / trial period, generate top , aggregate features for models for all customers, up to 50M lines per day (Python, docker)
o Dashboards analysis for model’s metrics or data analysis in general on Looker

• Data scientist:
o Model for predicting Churn (Tensorflow, keras) (65% accuracy)
o Model for predicting client’s conversion (from trial period to subscribed) (70% accuracy) and around 3-7% improvement in retention.
o Anomaly detection model for royalty’s fraud. (Clustering)
o Persona for clients , every new clients get predicted a “Persona” based on historical clients which is used to calculate a more accurate LTV / ARPU
and understand better what are the clusters of clients we have
(Example : majority of clients are X years old and like Jazz mostly and never download music) ,
currently used by marketing team for better mailing campaign. (Julia MLJ , Clustering)

iVactis

High Tech

Data Scientist / DevOps / Developer

Lyon

September 2018 - August 2019

Developed many micro services around data and Machine Learning with Docker / Kubernetes including :
- Machine learning Models (Scikit-learn, TensorFlow) - Report on data analysis.
- Manage databases from different sources (MSSQL, PSSQL) - API (Flask) - mobile app (Android in Kotlin)
- Real time pathfinding algorithm (Derived A* in python3)

AirVisual | The Air Quality Community

Environment

Scraper Developer

Beijing City, China

March 2017 - August 2017

Real time data mining from air quality related websites. Scrap data -> Filter -> clean -> add to database Also made minor Forecasting. With MeteorJS and CoffeeScript

Mereos

Education & E-learning

Lead Developer Chat bot + E-commerce

Lyon

October 2016 - March 2017

3 external recommendations

Check out Fabien's recommendations

Education