Mar 2016 – Jul 2018 · 2 yrs 4 mos · Toulouse Area, France

• Since 2018: Architect & core developer for OVH Kubernetes PaaS
• Development of the core API to deploy and manage multiple Kubernetes clusters, nodes and services
• Kubernetes and Docker expertise for internal needs
• Infrastructure as code and continuous deployment (Packer, Terraform, Openstack)
• 2017 - 2018: Architect & core developer for OVH Functions (serverless platform)
• Design of a distributed, resilient, scalable microservice architecture
• Asynchronous event-driven communication layer based on Apache Kafka
• Development of multiple core components in Go (Front-end API, Docker images builder, execution scheduler, Kafka trigger, runner agent on worker nodes, etc.)
• Operation and release cycle based on Jenkins, Docker, Ansible, Openstack and Terraform
• Focus on user experience and time-to-market
• Since 2017 SRE for OVH Queue (Kafka as a Service product)
• Service in production for 1.5+ years. 19 clusters, 90+ nodes, 500+ TB of
• storage disks. 1.5M messages per second average, with peaks at 3M
• Training sessions for new team members on the stack
• Continuous improvements on deployment, operations, monitoring and
• alerting
• Writing of post-mortems, focus on solutions for major outages and
• regular incidents
• 2016 - 2017 Architect & core developer for OVH Queue (Kafka as a Service
• product)
• Design of a multi-tenant Kafka as a Service solution
• Development of a performant TCP proxy in Go for multi-tenancy over Kafka protocol
• Development of an auth plugin in Scala, backed by Redis
• Development of an HTTP proxy to ease Kafka message production
• Development of several microservices to handle Kafka metrics forwarding, topic management, consumer lag monitoring, users and roles creation, synchronization between clusters, end-to-end integration tests, health-checks for the entire stack
• Development of the operations and release cycle, based on Docker and Ansible
• Kafka expertise for internal and external customers
• Evangelisation and talks about Apache Kafka

Sep 2013 – Feb 2016 · 2 yrs 5 mos · Toulouse Area, France

• Member of the Euclid Consortium System Team for 2.5 years
• International distributed context
• Contributions to the architecture design of a distributed spatial imagery data pipeline
• Contributions to the development of the core Java API to schedule pipelines and abstract inner infrastructure across heterogeneous computing centers
• Development of a prototype in Java with Apache HBase and Hadoop for scalable storage and computing of space sources catalogs
• Development of a prototype for product metadata storage and search engine based on Elasticsearch
• Development of tools for automatic checks and validation of scientific XML-based data model, automatic classes generation for C++ and Python
• Contributions to the development of an HPC job scheduler in Python
• Several talks and internal trainings presented during Euclid System Team meetings
• Member of Capgemini Big Data Group, facilitated training sessions based on:
• Machine Learning (Andrew Ng Coursera materials)
• Functionnal programming in Scala (Martin Odersky Coursera materials)
• Apache Spark (Berkeley edx materials)

Mar 2013 – Aug 2013 · 5 mos · Toulouse Area, France

• Dans le cadre du projet spatial Euclid du CNES (cartographie de la matière noire de l'univers), étude et développement d'une solution Hadoop/MapReduce de chargement, traitement et stockage des données récupérées depuis le satellite. Technologies liées: Hadoop, HBase, Pig, Dumbo, Java, Python.

Jan 2012 – Jul 2012 · 6 mos · Rennes

• Développement d'une application en Java permettant de traiter des mesures SPC de Pulltest de manière automatisée

Jun 2011 – Aug 2011 · 2 mos · Rennes

• Développement d'une application C++ & Visual Basic pour l'automatisation de mesures d'antennes couplées à un GPS