Scale
Operated 47 Kubernetes clusters across 22 cloud and on-prem regions.
Ownership
Owned SRE roadmaps with CTOs; led teams of senior engineers.
Breadth
10 years in production across AWS, Azure, GCP, Scaleway, and on-prem - ~50% on-prem at Kiln today.
Developer Experience
Platforms as products - developer self-service from commit to production.
Security
Security by design - workload identity, zero trust at scale, supply-chain hardening, secrets management.
AI Platforms
Shipping AI-assisted tooling and agentic workflows for platform teams at Kiln; drove AI/agentic adoption across engineering at Hublo.
Experience
-
Kiln
- Platform engineering & SRE for enterprise-grade digital-asset infrastructure.
- Building AI-assisted tooling and agentic workflows for the platform team.
-
Hublo
- Owned the SRE roadmap with the CTO; led a team of 3 senior engineers.
- Led the ECS→EKS migration and directed the AWS network foundation rebuild (VPC & IP-addressing design) to unblock multi-region growth.
- Drove adoption of AI tooling and agentic platforms across engineering.
-
SOPHiA GENETICS
- Operated 47 Kubernetes clusters across 22 regions (Azure + on-prem) for a global genomics SaaS.
- Migrated 15 TB of on-prem data to Azure with Data Factory.
- Coordinated distributed engineering teams across Europe.
-
Ippon Technologies
- Actia Automotive - software forge modernization, Azure landing zones, Kubernetes execution platform.
- Berger-Levrault - monolith→microservices replatforming: Istio, Tekton, ArgoCD, security-by-design; developer platform on K8s.
- Twelv - sovereign-cloud SaaS architecture on Scaleway Kubernetes.
Earlier: DevOps Engineer @ SKALE-5 (2021-22) · DevOps Architect @ RGPD Web (2019) · DevOps Architect & IT Manager @ Genesis Groupe (2018-19)
Selected Work
Hublo · Self-service Kubernetes platform · 2024-2026
From product spec to production in one commit
The problem
Shipping to production meant routing through the platform team. Every new service was a ticket, a wait, and a handful of inconsistent decisions about networking, observability, and security. That doesn't scale - not the infrastructure, and not the people maintaining it.
The approach
Before committing to a line of infrastructure, I interviewed the stakeholders and the platform's real customers - our developers - to build for an actual need, not an infrastructure trend. That product mindset is the difference between a platform people are forced to use and one they reach for.
What I led - across the stack
Developer paved path
Directed the design of a golden path that takes a service from product spec to production in a single commit. Observability, alerting, and security are part of the path, not an afterthought.
Kubernetes platform
Led a re-architecture of cluster operations around four non-negotiables I refused to trade off against each other: developer autonomy, performance, security, and cost.
AWS network foundation
Drove the rebuild of the networking layer (VPC and IP-addressing design) for scale - and for safe, secured interoperability with private partners.
~3 days → ~45 min
New service: zero to production
4 teams
Shipping autonomously
Zero
Platform tickets in the deploy path
Observability, alerting, and security ship on the paved path by default - reliability built in, not bolted on.
Stack Kubernetes · ArgoCD · Terraform · AWS (EKS, VPC, Transit Gateway) · Backstage · Prometheus · Grafana · Datadog · OpenTelemetry · NATS
Stack
Orchestration
Kubernetes · Crossplane · Helm · Kustomize
Clouds
AWS · Azure · GCP · Scaleway
IaC
Terraform · Pulumi · Crossplane · Ansible · Packer
GitOps & CI/CD
ArgoCD · Flux · GitLab CI · GitHub Actions · Tekton
Observability
Prometheus · Grafana · Datadog · OpenTelemetry
Networking
Cilium · Istio · Gateway API · Hub&Spoke · Network Policy
Security & Identity
SPIFFE · SOPS · Vault · Wiz · IdP · ZeroTrust
Messaging
Kafka · NATS
AI platforms
Agentic workflows · LLM infrastructure · MCP