Mohamed Hedi Ferjani

1- SmartOps/Ghosmod /Well Guardian/Anti -Coll : Data & Applications Drilling & Wells

Responsibilities :

▪ Acted as a transverse technical referent on several projects related to

exploration, drilling, and wells activities (Drilling & Wells)

Project 1 : Direct Feeding (Risk Analysis team): Migration of the trading metrics collection, normalization, and storage solution for financial products.

Responsibilities :

▪ Azure Migration: Implementation of the Azure EventHub streaming solution

▪ Dockerization and deployment of various chain integration services in Kubernetes

▪ Performance improvement and various services (multithreading, data processing parallelization, execution time improvement…Etc)

▪ Development of the OPAL solution: Exposure of "versioned" datalake data via Spark Java/Scala APIs

▪ Client support during chain tests: integration, quality, performance, and non-regression tests)

Project 2 : Decommissioning Exadata (North Star team): Removal of the Exadata component and migration of various data extractions to the datalake.

Responsibilities :

▪ Scoping and design of the migration solution

▪ Participation in various technical workshops

▪ Development of shared utility libraries

▪ Writing technical documentation

▪ Client support during the different testing phases

Responsibilities :

▪ Support developers from different FTs in upskilling on various topics related to quality and best practices: BDD, TDD, CI/CD, Microservices, Clean Architecture, Design Patterns…etc.)

▪ Monitoring, reviewing, and validating developers' learning on chosen topics (presentations, Katas…etc.)

Project 1 : mlBox: A web-based self-learning solution (Auto-Machine Learning) offering various features for data scientists to simplify and automate the creation of Machine Learning models through user-friendly interfaces.

Responsibilities :

As the technical lead of a team consisting of: 1 Data Engineer, 1 DevOps, and 2 Data Scientists and 1 intern, I was involved in:

▪ Scoping and design of a solution from scratch based on Artificial Intelligence and Machine Learning challenges.

▪ Implementation of a distributed computing solution for ML algorithms (Python, scikit-learn, Dask)

▪ Industrialization, refactoring, and optimization of data science development

(preprocessing, feature engineering, feature selection, discretization, modeling,

hyperparameter optimization, etc.).

▪ Design and development of various utility libraries (dask-utils, store-utils, test-utils, logging-utils, docker-utils)

▪ Exposure and REST API creation for the mlBox engine (Python, Tornado, MongoDB)

▪ Dockerization of the application's various components

▪ Definition of various CI/CD pipelines (GitLab, Nexus, Docker, Artifactory, Kubernetes, OpenStack)

▪ Support for production managers during deployment (technical documentation, scoping workshops, and presentations)

Project 2 : Score Habitat: A Machine Learning solution for calculating real estate credit propensity scores: This project aims to create ML models based on the aggregation and consolidation of multiple data sources to develop a model for scoring interested clients and helping them realize their acquisition projects.

Other Responsibilities :

▪ Leading "Evangelization" sessions and best practices workshops for data scientists and data engineers titled "Question Pour Un DE" (A Quiz for a DE).

▪ Internship supervisor: "MLBOX version PySpark"