In an HDS (Health Data Hosting) environment, conducting a study for changing the API gateway to expose microservices

• Development of Mocks to limit external partner calls for performance testing
• Benchmarking of 3 API gateway solutions (Kong / Traefik / APISIX)
• Building route profiles (circuit breaker, security, JWT management, rate limiting)
• Implementation of autoscaling with Karpenter and robustness testing

AWS, EKS, Karpenter, Traefik, FluxCD

Within an EDF subsidiary, redesign of an IoT architecture for collecting photovoltaic supervision data. Ingestion of 57M events per month.

• Microservices architecture design
• Definition of business domains and technical foundations
• Development of a technical roadmap
• Writing multi-stage architecture diagrams
• Writing specifications and ADRs
• Setting up AWS infrastructure

AWS, Fargate, Cognito, SQS, Pulumi

Supporting the team on the LLMOPS part on K8s Rancher and Azure

• Building on-premises k8S infrastructure with A100 GPUs
• Implementing LLM monitoring tools
• Creating a roadmap for cloud > on-premise migration
• Guiding the team in GenAI technology choices
• Setting up experimentation pipelines with Mlflow

Azure, Rancher, Langfuse, Promptfoo, liteLLM, Openlit

At Kazaplan, I was in charge of cross-functional architecture for the entire project as the Infrastructure Engineering Manager.

• Team building (recruitment, planning, roadmap)
• Technical solution arbitration
• Scrummaster / agile rituals
• Creating a shared roadmap and engaging key players
• Setting objectives

Notable technical achievements:

• Rendering with Blender on GPU with autoscale
• Implementation of testing tools on kubernetes
• Deployment of AI Engineer architecture

GCP, Kubernetes, Pulumi, Gitlab CI/CD, Traefik, Argo CD...

Technical side:

• Microservice design
• Monitoring 
• Security
• Performance
• Technical concept injection (sse, sso, lambda, cloud, kubernetes)
• Data model redesign

Operational side:

• Lead coach (posture, role, responsibility)
• Cross-functional workshops (feature flipping, migration)
• Development guidelines
• Tool training (sonarqube / datadog)
• Improvement management (kpi / board / measurement)
• Agile transformation support (scrum / accelerate)

A few example projects:

• Workshop on lead dev posture
• Micro-service design (Anti Corruption Layer pattern)
• Development stack redesign
• Technical roadmap construction
• Data model redesign
• Training teams in technical documentation (code / architecture)

GCP, Kubernetes, Terraform, Gitlab CI/CD, Node, Php

Guided teams to set their architectures on Azure and AKS, lead workshops, and benchmarked Kubernetes platforms.

AKS, Azure, Terraform, Gitlab CI/CD, Dotnet Core, Keyvault

• Kubernetes architecture training for the tech team
• Schedule and roadmap set up
• Data audit to validate the project's suitability for requirements
• Workshops for developers on microservice architecture
• Coach the Ops staff on Kubernetes implementation best practices and security.

• Coaching the team on Kubernetes implementation architecture in GCP in a global CDN context
• CD implementation with argo CD
• Full google platform infrastructure (gke / cdn / ingress / etc)

GCP, Kubernetes, Terraform, Gitlab CI/CD, Argo CD

Played a part in the migration of an IAM platform based on a forgerock solution.

1. Building a new platform

    - Construction of project technical roadmap to set deadlines

    - Refactoring of helm charts (statefulset, cronjob, etc...)

    - Search for cloud backup and secret management solution

    - Provide secure docker images alpine base (stack httpd - java - tomcat - nginx)

    - Performance benchmark with jmeter scenario (SAML / OIDC)

    - Installation and configuration of datadog (log / metric /APM)

    - Proposal for kubernetes architecture on AWS (EKS, S3, secret manager.... )

    - Security audit management (pentest blackbox / whitebox)

2. Development

    - Backoffice development with python Flask / VueJS to administor the new platform's OIDC SAML client configurations.

    - 30% MCO saved

    - Deployment configuration in 5 minutes rather than 5 days (automating IS connections)

    - Configuration deployment on LDAP directories, identity federation and web services

    - Interface with forgerock and SNCF APIs

    - UX / UI design, data modeling, secure routes etc...

AWS, Flask, Vuejs

Migration from identity federation architecture to a Kubernetes stack

AWS, Kubernetes, Am Forgerock, Jenkins

Built Kubernetes service offering for SNCF on AWS

AWS, Azure Kubernetes, Python, Vuejs

Working on M6 Web data lake with CDH 5.11 Hadoop cluster to run A/B tests.

Projetcs :

• Data-lake migration (Hive table, HDFS data) with quality data verification
• Setup and installation of continuous integration with Jenkins, Sonarqube, Artifactory and Docker
• Centralisation of all projects configurations into json file (backup, quota, retention, access right, ....)
  

Other tasks

• Deployment automation in clusters with Ansible (add new users / projects / quota)
• Development, orchestration and scheduling of jobs using Python & Airflow ( coding specials plugins)
• Setup Data-lake cleanup with Java code (delete HDFS data, clean hive partition)
• Build Docker containers for developper
• Setup backup data in S3
• Setup EC2 machine with VPC on AWS
• Tuning Hadoop Cluster to improve performance
• POC with Apache Presto to get fast database over hive tables.
• Data Science tool configuration (Hue / Jupyter / SuperSet)
• Python script developement to help collegues in various use cases.
• Analysis and resolution of incidents
• Support to developers, data analysts and data engineers
• Migration code from Spark 1.6 to Spark 2 to enhance performance

Hadoop, Ansible, Python