Hello 👋

I’m a senior Software Engineer building data infrastructure at Clever Cloud. I’ve built and operated a variety of stateful distributed systems throughout my career.

I’m interested in distributed systems, data stores, understanding how things works under the hood. I enjoy being part of open-source communities, through talks, posts, and contributions to:

I am also maintaining some projects such as:

On my free time, I am giving a hand to local events, such as the local GDG/JUG FinistDevs, Devoxx4Kids and I am also a teaching assistant in my former Engineer School.

I co-founded in 2017 HelloExoWorld, an initiative to search for exoplanets using Warp10, a time-series platform.

Work

2021 to now: Senior Software Engineer @Clever Cloud

I am working around data infrastructure, to offer new database-oriented products 🚀

2020 to 2021: Technical Leader @OVHcloud - Managed Kubernetes

I was working on the managed Kubernetes product. I was involved in:

  • Improving ETCD’s scalability and OpenStack CSI,
  • Self-healing customers control-planes,
  • Go mentoring,
  • On-call duty.

2019 to 2020: Technical Leader @OVHcloud - ioStream

I was working on the underlying infrastructure of IO oriented products, including ioStream, a geo-replicated, managed topic-as-a-service product built using Apache Pulsar.

During that time, I worked around adding Kafka’s protocol to Pulsar.

We launched a beta mid-2019, and the project has been shutdown mid-2020.

2016 to 2019: Infrastructure Engineer @OVH

I worked on Metrics Data Platform. We are using Warp10 with friendly Apache softwares such as Hbase, Hadoop, Zookeeper, Kafka and Flink to handle all OVH’s metrics-based monitoring, which represent around 432 billions of measurements per day.

I have taken part of most of Metrics development, from internal management to Ingress/Egress translation part. I also worked on the implementation and deployment of a distributed and scalable alerting system using Apache Flink.

I was using Flink, HBase, Hadoop, Kafka, Ansible, Go, Rust, Java, Linux, WarpScript on a daily basis.

I was on-call duty on more than 700 servers, including:

  • 3 Warp10 fully distributed clusters, including one that handling 1.5 millions datapoints per second
  • 3 Hadoop clusters, including a 250 nodes Hadoop cluster running a 75k regions Hbase table
  • Various Apache technologies, such as Kafka, Zookeeper and Flink

I gave training and support for both external and internal client of Metrics, including WarpScript.

During that time, I contributed to Apache Flink and HBase.