Ali TBER

Having fun with : - Forecasting - Machine learning Engineering - MLOPS

Leading projects to deploy tools to optimize the supply chain thanks to accurate forecasting with machine learning. Working with the pricing and promotion team to build a global product helping the business. Technical stack: Python, Sql, Pyspark, Jenkins, MLflow, Pytest, Git, Docker, AWS Contributor in the migration from AWS to GCP. (ML Engineering)

Working on Data science for business projects. Consulting with all the positive aspects. • Implementation on Azure of Advanced forecasting algorithms to predict the demand for a media group. In charge of the system design, development, and industrialization of the global algorithm • Creation of a big data pipeline with digital data in order to centralize the digital customer journey of automobile customers, from the website to the dealers. More than 4To of data linked to a global app which allows media optimization • Innovation project on natural language processing. Focusing on transformers for text classification and text generation • Development of a customer lifetime value for the aftersales customers of an automobile manufacturer Technical stack: A lot of Python(Pandas, Sklearn, Pytorch), SQL, Pyspark, Databricks, Azure cloud, Git, MLFlow

• I joined the customer knowledge team with more than 10 experienced statisticians : - Development of a tool using R for automating the different scoring used. Churn/propensity score - Delivery of a tool that outputs an explainable model and a score with the different non-linear models. - Deep learning for image analysis (1month)

In collaboration with the Data analytics team : • Development of a predictive tool for a bank in order to classify emails into 8 categories : - Benchmark of different libraries to do natural language processing. Test of the performance and usefulness of each library (Nltk, Spacy, Gensim) - Work on predictive models and many feature engineering optimization before (sci-kit learn, Keras) • Development of a web application prototype taking an email and outputting a prediction (Flask, Python)