Idriss Brahimi Email & Phone Number

Paris, Fr

Ville De Paris, Île-De-France, France

- Design and implementation of data pipelines for real-time processing for display onstore vendors' tablets- Management of architecture migration: transition from planned to event-driven processing via Vertex pipeline and DBT- SQL optimization to reduce analytical query response time- Industrialization of containerized Machine learning models (via BQML)- Monitoring / Alerting management of the pipelineTech Stack: GCP (DBT, Vertex, Docker, Bigquery, Terraform, Composer, IAM, BQML, Storage, Cloud Logging, Cloud Monitoring, Cloud Run, Pub/Sub, Cloud Function, GSR)

Ville De Paris, Île-De-France, France

- Framing of the functional need- Management of payment inflows via the Spark framework- Aggregation of banking transactions by the median (Robust Estimator)- Log-normal transformation of transaction medians and Kurtosis optimization- Use of a Gaussian Mixture Model (GMM) type classification model with optimizationof the number of clusters by the weighted BIC criterion.- General change in the behavior of all clients over time (Data Drift): Re-training of the modelTech Stack: PySpark

Région De Paris, France

Implementation of a voice module to make the input method easy in a application who track the path of a food. - Development of the interface front/back of the voice module - Training of a DialogFlow model and development of a script to retrieve the text of the voice module and send him to DialogFlow model. Participation on the “E-commerce Prediction” use case, which concerns the creation of a model for predicting the number of orders for all Carrefour Drives - Improved of an existed prediction model based on Prophet library of Facebook - Creation of syntetical historical data for a better predictionParticipation on the "Optimisation of product's assortment" use case, which concerns the creation of a recommendation model to improve the product assortment of all Carrefour Citys to increase their sales - Implementation of a dashboard in a tool named SuperSet - Implementation of metrics to judge to fiability of a modelLibraries : Pandas, Prophet, Flask, DialogFlow, scikit-learn

Région De Paris, France

Implementation of a recommendation system of job's offer in PeopleFlare, a platform who put in a relationship freelances and companies (Natural Language Processing)Language : PythonLibraries : Gensim, Numpy, ScipyCreation of a path tracker of all the users on Daylindo, a platform who help to adopt good habits on different subject of the daily life (sport, aestetic, dietetic, …)Language : R