Freelancer profile translated to English.

Description

I design and deploy complex data platforms in constrained enterprise environments — regulated, air-gapped, multi-team.

My domain: IaC Ansible/Terraform, Kubernetes/OpenShift orchestration, SSO/OIDC integration, and large-scale pipelines. Most recently, full deployment of a 6-node Dataiku DSS platform (zero touch, 15 Ansible roles) with H100/H200 GPUs, advanced MIG configurations, and distributed Spark at a major energy company.

I integrate AI tools into the development cycle — not as a gimmick, but to reduce the time between design and operational delivery.

Industry field of expertise

Languages

French
Native or bilingual
English
Native or bilingual
Arabic
Native or bilingual

Workplace preferences

Remote only

Primarily works remotely

EDF SA
Data Platform Engineer
ENERGY AND UTILITIES
December 2025 - Today (8 months)
Paris, France
Design and full IaC deployment of a Dataiku DSS 14.4 platform with 6 production nodes (Design, Automation, Govern, Deployer) in an isolated enterprise environment (no internet access, corporate proxy, dual VLAN):
Development of 15 Ansible roles and 1 custom module orchestrating a 15-phase deployment pipeline with full rollback in a single command — zero manual configuration.
In-depth mastery of Dataiku DSS: advanced configuration via API (SSO, Spark, connections, governance),
Multi-layer permission architecture by business entity: node access (Keycloak), granular permissions (29 flags per group), ACL per connection, and rights per project — with isolation of project folders and traceability of compute costs (chargeback).
Development of a complete administration web app (Vue 3/TypeScript + Flask): user/group/entity management, unit and bulk CSV onboarding, Keycloak auto-logout upon rights modification.
Setup of containerized execution profiles (CPU, GPU MIG 1g/3g/7g) and optimized multi-stage image builds (GPU/CUDA, OpenCV, Spark, Python/R) for 50+ data scientists.
Kubernetes (OpenShift) integration with NVIDIA H100 GPUs and distributed Spark, SSO/OIDC authentication via Keycloak with automatic provisioning and LDAP/MFA.
Resolution of complex infrastructure incidents: cross-VLAN pod-VM connectivity, 3-layer proxy (Shell/Java/Application), rootless Podman on locked VMs.
OpenShift Dataiku Podman Keycloak Architecture
EDF SA
Data/DevOps Engineer
ENERGY AND UTILITIES
May 2022 - December 2025 (3 years and 7 months)
Paris, France
Building and managing the data platform for EDF's data science teams, from infrastructure to use case delivery:

DevOps & Platform (2022 – 2025):
Deployment and management of a Docker Swarm cluster (3 compute nodes) and automation of the complete data platform (Airflow 2.10→3.1, Traefik, Trino) on OpenShift and VMs
Build of optimized, multi-stage, and secure Docker images (GPU/OpenCV, data science dependencies) for all teams
Introduction and migration of Python packaging tools: Poetry → uv/ruff, standardization of development practices
Migration of Airflow DAGs during the 2.10 → 3.1 transition

Data Engineering (2022 – 2025):
Migration of data science projects from Spark/Hadoop to Python/Polars/Oracle SQL, reducing resource costs by over 40%.
Design and development of 12+ REST APIs to expose and industrialize data science use cases (FastAPI/Dash/Traefik)
Development of connectors and internal data ingestion library with unified API (Python/Polars/Pandas/SQLAlchemy)
Training data scientists on best engineering and optimization practices

R&D Data Pipelines (2024 – 2025):

Processing terabytes of multimodal scientific data (imaging, sequencing, structured data) for an R&D decision support system
Optimization and parallelization of large-scale processing pipelines on AWS (EMR, S3, ECR, Airflow, Terraform)
Refactoring of legacy pipelines to Apache Airflow, monitoring with Grafana/Prometheus
Complete multimodal anonymization (images, text, audio) GDPR/CNIL compliant.
Python Polars Docker SQL Agentic AI
Groupe Crédit Agricole (AE)
Data & AI Engineer
BANKING AND INSURANCE
October 2019 - April 2022 (2 years and 7 months)
Paris, France
Within Crédit Agricole's Datalab, developing and industrializing Data & AI solutions for banking operations:

Design and implementation of REST APIs for fraud detection and OCR data extraction, accelerating the KYC process for Crédit Agricole's regional banks across France (Python/FastAPI/Tesseract)
Contribution to the development and deployment of an internal distributed AutoML solution leveraging Spark and H2O.ai
Optimization of Spark job performance and debugging within a Cloudera cluster, maximizing operational efficiency
Creation of a PySpark Template to automate the packaging and deployment of Spark jobs on isolated Cloudera platforms (Shell, PEX, GitLab CI/CD)
Design and facilitation of MLOps, industrialization, and distributed systems workshops
Testing of GPU libraries (PySpark/spark-rapids, cuDF, H2O Sparkling Water) on a GPU Kubernetes cluster.
PySpark Python Cloudera MLOps Kubernetes