Assess
DevOps and platform audit
A review of how code gets from commit to production, with a written report and a ranked roadmap.
- CI/CD, Kubernetes, IaC and cloud setup
- Reliability, security and cost findings
Training
Programmes for teams that run real systems
Private sessions for your team, on site or remote. Each course is mostly labs, uses course material written in-house, and gets adjusted to your tools, your cloud and your level before day one.
DevOps & Cloud Native
Containers to productionDocker, from first container to production
Images, volumes, networking, Compose, multi-stage builds and image hardening.
2–3 daysKubernetes for developers
Deploy, configure and debug applications on Kubernetes without becoming a cluster admin.
3 daysKubernetes administration
Running the cluster: upgrades, backups, RBAC, networking and storage, on a managed cluster or one you built yourself.
3–5 daysKubernetes on the SUSE stack
Rancher to run the fleet, RKE2 and K3s to build it, then Fleet, Longhorn and NeuVector for delivery, storage and security.
3 daysHelm and GitOps with Argo CD
Charts that survive upgrades and rollbacks, then a Git-driven delivery pipeline you understand end to end.
2 daysCI/CD with GitLab
Pipelines, runners, environments, container builds and deployments to Kubernetes.
1–3 daysAnsible, from playbooks to production
Playbooks, roles and collections, then the day-to-day automation of your servers and your deployments.
2–5 daysAWX and Ansible Automation Platform
Ansible run as a shared platform: inventories, workflows, execution environments and who is allowed to run what.
3 daysObservability and SRE
Prometheus, Grafana and alerting, plus SLOs, error budgets and incident practice from SRE Foundation.
2 daysDevSecOps in practice
Supply-chain security, vulnerability scanning, image and cluster hardening, NeuVector.
2 daysLinux for cloud native engineers
The system under the containers: processes, systemd, networking, permissions, storage and diagnosing a box in production.
2–3 days
Generative AI & AI engineering
Prompts to shipped agentsGenerative AI and prompt engineering
How LLMs work, what they get wrong, and prompts designed, tested and versioned like code.
2–3 daysBuilding LLM applications: RAG, tools and MCP
Retrieval pipelines, function calling, MCP servers and clients with FastMCP and LangChain.
3 daysAgentic development and multi-agent systems
Agents, orchestration, guardrails and human approval, from prototype to something you can operate.
3–5 daysLocal AI with Ollama
Sizing hardware, serving models on your own machines, Modelfiles, QLoRA fine-tuning, private agents.
3–5 daysGitHub Copilot for engineering teams
Copilot and Copilot Chat on a real codebase: what to ask for, review habits that catch its mistakes, security and licensing, and measuring whether it helps.
2–3 daysAI agents for business teams
Claude Cowork and Copilot Studio: skills, tool connections, deliverables and governance, without code.
2–3 daysClaude Code on your codebase
Driving a coding agent on real work: project rules and skills, MCP tools, subagents running in parallel, plan and review, plus token cost and safety.
3 days
Foundations & development
Git to microservicesGit, properly, in a day
Working directory, staging area and repository, branching that a team can live with, and how to undo anything you did.
1 dayPython from zero to microservices
Python for engineers who ship services: APIs, packaging, tests, containers and deployment.
3–5 days
Executive briefings
Decide before you investCloud native for decision-makers
What containers, Kubernetes and managed cloud actually change for cost, hiring and delivery time — and which parts are not worth buying.
2–3 daysThe technology shift, without the hype
Where AI, cloud and automation are genuinely heading, what that means for your teams, and what is worth deciding this quarter.
2–3 days
In-house, private
On site or remote
English or French
Custom programmes
Consulting
Hands-on help, then a team that can carry on alone
Short, scoped engagements. The work is done with your engineers, not handed over as a black box, and it ends with documentation and a session so your team owns what was built.
Build
Kubernetes and GitOps platform
Clusters, Helm charts and Argo CD pipelines designed with your team and running in your environment.
- Cloud or on-prem, RKE2 and K3s included
- Promotion flow, rollbacks, secrets
Operate
Observability and reliability
Metrics, dashboards and alerts that point at real problems, plus SLOs the team agrees to defend.
- Prometheus, Grafana, alert tuning
- SLOs, runbooks, incident reviews
Automate
Infrastructure automation
Ansible roles and an AWX or Automation Platform setup, so server changes stop being manual and other teams can self-serve.
- Inventories, roles, execution environments
- Workflows other teams can run themselves
Secure
Supply chain and DevSecOps
Hardening the path from commit to cluster, with checks that fail builds for a reason the team can act on.
- Image and cluster hardening, scanning, SBOMs
- Policy gates in CI, secrets handling
Build
Generative AI applications
RAG assistants, agents and MCP servers connected to your data and tools, with evaluation from the start.
- Retrieval quality and answer evaluation
- Tool access, approvals and audit trail
Host
Private and self-hosted AI
Models running on your own infrastructure when data cannot leave the building.
- Hardware sizing and model choice
- Serving, access control, monitoring
Adopt
AI adoption for engineering teams
Rolling out coding assistants and agents with rules, guardrails and a way to measure whether they help.
- Team conventions, shared skills and rules
- Security, data policy, cost tracking
How it works
From first call to delivery
Most training requests get a proposal within a week. Consulting starts with a fixed-scope first phase, so you can judge the work before committing to more.
Scoping call
30 minutes on your team, your stack and what should change after the training or engagement.
Written proposal
Programme or scope, format, dates and a fixed price. Existing programmes are adapted to your tools.
Delivery and follow-up
Sessions or project work, then the material, labs and documentation stay with your team.
About
Behind eralabs: Aymen El Amri
eralabs is the training and consulting practice of Aymen El Amri, an engineer and instructor based in Paris. Every session and engagement is delivered by him directly.
He has trained engineers since 2017, written more than twenty technical books on Docker, Kubernetes, GitOps, Ansible, prompt engineering, MCP and local AI, and founded FAUN.dev, a developer community of 100,000 engineers.
- RecognitionTechBeacon DevOps 100
- AmbassadorDevOps Institute, France, since 2019
- eralabsCIO Applications: top 10 DevOps consulting in Europe
- BooksPublished by FAUN.dev, Packt, O'Reilly library, Leanpub; translated into five languages
- SpeakingDevOps India Summit, DevOpsDays Cuba, École 42, DevOps Institute SKILup Day