Microsoft AzureAKS, networking, identity, Key Vault and data
Tools selected for purpose, not popularity
Technologies I work with today
I select technology around the problem, the team's operating capability and long-term maintenance cost. My work spans public cloud and private infrastructure — from Azure, Kubernetes, Terraform and GitOps to physical servers, private cloud, storage, monitoring and networking.
01
Cloud and containers
Public cloud services and modern application platforms.
Kubernetes / AKSPlatforms, workloads, networking and operations
TerraformRepeatable infrastructure and reviewable plans
DockerContainers, local development and self-hosting
02
On-premises and private infrastructure
Owned hardware, virtualisation, storage and monitoring as one operable system.
PuppetCentral, repeatable configuration across physical and virtual hosts
OpenNebulaPrivate cloud, virtualisation and compute capacity management
CephDistributed storage, performance, availability and recovery
ZabbixMonitoring, discovery, alerting and capacity signals
Linux / DebianServer operations, network services, hardening and automation
Windows Server / VMwareAD, domain services, SQL, virtualisation and enterprise operations
AI INFRASTRUCTURE
I also operate specialised machines with large memory configurations and multiple GPUs. The work spans hardware sizing, power and cooling, storage and networking, drivers and model runtimes, monitoring and capacity planning.
03
Delivery and GitOps
Safe changes from repository to runtime.
Argo CDDeclarative delivery and drift control
GitHub ActionsCI/CD, OIDC and automated controls
GitLab CI/CDPipelines, runners and security scans
HelmPackaging and environment overlays
04
Security, data and operations
The layer that determines reliability.
Envoy GatewayHTTP routing, policies and internal services
PostgreSQLManaged data, migrations and permissions
PrometheusMetrics, alerting and operational signals
GrafanaDashboards and shared operational context
05
AI-assisted engineering
Agents as a practical part of the workflow, not a replacement for engineering judgement.
OpenAI CodexMy primary tool today for analysis, implementation and change verification
Agent workflowsParallel research, cross-repository work, testing and operational automation
ClaudePrevious experience with analysis, solution design and large-context work
GitHub CopilotCode completion and fast assistance with everyday work inside the editor
LOCAL AI LAB
I also experiment with smaller local models on my gaming PC, including models from the Qwen family. I compare their practical usefulness, response quality, speed and hardware requirements, along with the privacy benefits of running workloads without sending data to an external service.
06
Home lab
Enterprise principles at a practical everyday scale.
Home AssistantAutomation, heating, lights, media and garden
MikroTikRouting, NAT, segmentation and VPN
Nginx Proxy ManagerTLS and secure exposure of local services
Self-hosted servicesLocal websites, automation and operation independent of user login
07
Working environments selected for purpose
No single platform is the best answer to every problem.
Bazzite on my gaming PCA Linux environment that gives me a fast, stable and exceptionally polished gaming system
macOS for everyday workMy iMac and MacBook combine a Unix foundation, security, performance and a refined user experience
Windows and the Microsoft ecosystemA strong choice for enterprise environments, central management, identity and broad compatibility
Context before preferenceI assess platforms around their purpose, operating model and the people who will use them
Broader experience
Tools change; principles remain
My experience also includes AWS, GCP, Oracle Cloud, Ansible, HAProxy, Vault, ELK/Wazuh, Fortigate, Palo Alto, VMware and operating enterprise Windows and Linux environments. This is not a logo collection, but tooling used in real operations that informs how I select an appropriate solution.