Description

With deep understanding of data science and machine learning and 9 years' corresponding experience (4 years in R&D, 2 years at IBM, 3 years at Sodexo), I have rich experience of each stage of Data and AI project (POC, Scoping, MVM, MVP, Industrialization) in different sectors (bank, restaurant services and facilities management, retails, energy, aeronautique, pharmacie). My various project oriented and client-facing working experience makes me not only able to fullfill technical work but also enable me to play the role as project manager and data translator. I hold a PhD degree with qualified competence and skills in innovation and R&D work(four 1st author publications on top world class scientific journals). I speak fluently french, english and chinese.

Languages

Chinese
Native or bilingual
English
Native or bilingual
French
Native or bilingual

Workplace preferences

Remote only

Primarily works remotely

SODEXO
Senior data scientist
RESTAURANTS AND FOOD SERVICE
June 2020 - November 2022 (2 years and 5 months)
Issy-les-Moulineaux, France
Lead data scientist on retail, pricing projects:
- Dynamic pricing for corporate France, US
- New food model of digital transformation in China
- AI asset initiation and developement

Participate the whole project life cycle: Project framing, MVM(minimum viable model), MVP(minimum viable product), Industrialization
In charge of conception and on boarding of data driven solution (BI, AI) to address business needs and pain point
Coach data scientists resources on retail, pricing projects

Main contributor of NLP asset of Sodexo Data Factory:
- Standard pipeline for NLP tasks include sentiment analysis, topics extraction, topics classification (Bert, GPT-3 etc)
Pricing NLP Fuzzy Matching segmentation MLOps Python Microsoft Azure
IBM France
Data Scientist & Machine Learning Engineer
TECH
June 2018 - June 2020 (2 years)
Bois-Colombes, France
1. Airbus Skywise project: predictive maintenance and anomaly detection by Machine Learning.

- Creation of data pipeline: massive timeseries data treatment and analysis by using Pyspark
- Creation of machine learning pipeline: models (deep learning: Autoencoder) development and hyper-parameters tuning (Bayesian Optimization)
- Deployment of data pipeline and machine learning pipeline for real time prediction to provide KPI in dashboard for anomaly detection

Technical environment: Pyspark, Tensorflow, Keras, Amazon AWS, Palantir

2. Ipsen & Roche: Prediction of adverse event of medications for Pancreas cancer; Prediction of treatment switch in diabetes type-2 patient journey.

- Defining hypothesis and cohort
- Feature engineering: cleaning data and creation of data model
- Advanced analytics: demographics analysis
- Creation of machine learning pipeline: deep learning model (LSTM) to predict time series event (adverse event); Cox analysis model (diabetes)
- Creating causal inference model for causality analysis and for eliminating bias in data

Technical environment: Jupyter notebook, Pycharm, Tensorflow, Keras, IBM explorys, SQL, Docker

3. POC for Nissan: Artificial intelligence solution for guided diagnostic of vehicles (Natural Language Processing NLP).

- Transformation of unstructured data (customer complains e-ticket) to structured data (NLP text analysis)
- Feature engineering: cleaning and standardising data
- Machine learning modelling: creating prediction models to precisely classify e-ticket.

Technical environment: Jupyter notebook

4. Compliance Watson for financial services of Crédit Mutuel: fraud detection and investigation of abnormal transactions against fraud and terrorist activities; customers retention.

-Feature engineering: cleaning, generation and exploitation of data for investigation
-Advanced analytics: data analysis
-Data modelling and AI modelling: creation of data model and machine learning prediction models.
Technical environment: Jupyter Notebook, SPSS, SQL, Watson Studio
Python Deep Learning Data Analysis Data visualisation AWS Microsoft Azure
Ecole Centrale de Lyon
Researcher
RESEARCH
June 2017 - May 2018 (11 months)
Écully, France
Conception and development of advanced model to predict high frequency acoustic and vibration effect in automobile and airplane by using historical data.

In charge of coordinating two R&D teams between Ecole Centrale de Lyon and Insa Lyon.
Innovation Data Modeling Mathématiques Algorithmes LaTeX