I break things professionally (before bad actors do), then automate the fixes. Lead SRE Security Engineer who believes infrastructure should be both bulletproof and elegant.

Currently obsessed with AI agents and making them play nice with security constraints.

I'm the person who says "mais non, we can't just YOLO to production" and then shows you how to do it safely with zero-trust architecture and proper secrets management. Think of me as the friendly neighborhood security guy who also happens to love Kubernetes chaos engineering.

These days? Building agentic solutions that don't compromise on security. Because AI agents are cool, but insecure AI agents are a nightmare waiting to happen.

Making Things Secure 🛡️

• Infrastructure hardening (because "it works on my machine" isn't a security strategy)
• Secrets management with hardware keys (SOPS, Vault, and my beloved Yubikeys)
• Zero Trust - because trust is for friends, not for networks
• Security automation - if I have to do it twice, it gets a script
• Compliance things (SOC2, ISO 27001, CISSP in progress) - the necessary evil

SRE Things ⚙️

• Kubernetes wrangling (because someone has to tame these clusters)
• Azure & GCP infrastructure (certified in Azure, GCP veteran)
• IaC with Terraform - because clicking in consoles is so 2015
• Making things observable (Prometheus, Grafana, ELK, Cloud Monitoring)
• On-call duty - my love-hate relationship with PagerDuty

AI & Agentic Solutions 🤖

• Building autonomous agents that actually work in production
• Securing AI/ML pipelines (because model poisoning is a real thing)
• LLM integration with enterprise security constraints
• Agent orchestration and workflow automation
• AI-powered security monitoring and response

Security Arsenal

Yubikey | SOPS | Vault | Falco | Trivy | OPA | Kyverno

Infrastructure

Kubernetes | Docker | Terraform | Ansible | ArgoCD | Helm

Cloud ☁️

Azure (certified) | GCP (love it) | Multi-cloud warrior

AI/Agent Stack 🤖

Claude | OpenAI API | LangChain | AutoGen | CrewAI
Prompt Engineering | RAG | Function Calling | Agent Orchestration

Languages

Go | Python | Bash | Ruby | HCL | YAML (yes, YAML is a language, fight me)

SOPS + Yubikey PIV

Because storing secrets in plaintext is a crime against humanity. This setup uses hardware-backed Yubikey PIV keys with physical touch requirement - if malware tries to decrypt your secrets, your Yubikey LED blinks and you know something's fishy.

C'est magnifique - your secrets are safe, your team is happy, and attackers are frustrated. Exactement what we want.

Agentic Solutions (Currently Building)

Working on autonomous agent systems that can:

• Self-heal infrastructure issues before humans wake up
• Analyze security logs and triage incidents automatically
• Orchestrate complex workflows across multiple tools
• Learn from production patterns (without breaking things)

The trick? Making agents reliable enough for production while keeping them on a tight security leash.

"Security without usability is just expensive theater. Good security makes the secure path the easy path."

"AI agents are like interns - powerful when properly supervised, chaos when left unsupervised."

Also:

• Automate everything (humans are terrible at repetitive tasks, agents are great at them)
• Monitor everything (you can't improve what you don't measure)
• Break things in staging (so they don't break in prod)
• Document everything (future-you will thank present-you)
• Trust but verify (especially with AI outputs)

• 🤖 Agentic workflows for security operations (current obsession)
• 🔐 Hardware security keys & zero-trust implementations (obsessed with Yubikeys)
• 🧠 AI-powered incident response (teaching agents to think like SREs)
• ☸️ Kubernetes runtime security (Falco is my spirit animal)
• 🔄 GitOps security workflows (because Git is the source of truth)
• 🛡️ Securing AI/ML pipelines (because someone has to)

Certified:

• ☁️ Microsoft Azure (multiple certifications)
• 🌩️ Google Cloud Platform (battle-tested veteran)
• ☸️ Kubernetes practitioner (learned the hard way, in production)
• 🚀 SRE principles (battle-tested with real incidents)

In Progress:

• 🔐 CISSP (the long one, but worth it)
• 🤖 AI/ML Security specialization
• 📚 Continuous learning (because tech moves fast)

• GitHub: right here, you found me
• Email: [email protected]
• Best way: open an issue on my repos (I'm more responsive than on email)

• I have more Yubikeys than most people have regular keys
• My AI agents have better error handling than my first production code
• I speak fluent YAML (unfortunately)
• Coffee-to-code ratio: 1:1
• Currently training AI agents to understand French humor (work in progress)
• "Have you tried turning it off and on again?" is a valid security strategy

I believe AI/LLMs are transformative for security and operations work, but only if we:

• Keep humans in the loop for critical decisions
• Build proper guardrails and security controls
• Test agents like we test code (because they are code)
• Monitor agent behavior like we monitor systems
• Accept that agents will fail (and plan for it)

The future is agentic, but it better be secure too.

"The S in IoT stands for Security" 😏

"The A in AI stands for... well, Agents now, apparently" 🤖

PS: If you're here from a security audit, yes, I practice what I preach. Check out my repos - actual working code, not just slides.

PPS: My AI agents are properly sandboxed and monitored. Promise.

Bisous! 💜