I take on scoped builds and migrations for engineering teams, covering Linux, AWS, Kubernetes, observability and automation.
Current role: Principal SRE in regulated fintech. Based in Lithuania, working European hours.
Examples from employed roles. Company and customer details are omitted.
At a previous employer, I led adoption of EKS Anywhere for low-latency exchange software being moved from a decades-old distributed bare-metal architecture. The software serves exchange operators across several countries. The environment was fully air-gapped and on premises. Preserving its latency characteristics required CPU core pinning and Solarflare NIC kernel-bypass networking for workloads running in Kubernetes.
In my current Principal SRE role, I lead observability work across the organization. My remit includes introducing SLOs and error budgets, driving OpenTelemetry adoption, and providing standard collection paths for logs, metrics and traces from cloud VMs and Kubernetes. Grafana Cloud, Grafana Kubernetes Monitoring and OpenTelemetry Operator auto-instrumentation reduce the setup required from application teams. I also led migrations from VictoriaMetrics to Grafana Cloud and moved Grafana Cloud stacks from GCP to AWS to use PrivateLink. Together, those changes cut estimated annual platform and data-transfer costs by about US$250k.
Most projects cross more than one of these.
On-premises and cloud clusters, networking, ingress, secrets, storage, backups, upgrades and handover. EKS, EKS Anywhere, k3s, Cilium and Argo CD where they fit the job.
AWS architecture, Linux systems, account and network design, compute, storage and application hosting. This also covers work that does not need Kubernetes.
Telemetry architecture, OpenTelemetry adoption, Grafana and Prometheus stacks, SLOs, alerting, cardinality control and cost reduction.
Terraform, Ansible, GitOps, delivery pipelines, provider migrations and small tools in Go or Python. This includes moving established systems without hiding operational constraints.
A short call to understand the system, constraints and intended result. If the work is outside my experience or availability, I will say so.
The proposal states the work, deliverables, dependencies, exclusions and commercial terms.
I carry out the agreed work, document what changed and walk the team through anything it will own afterwards.
I also keep limited room for ongoing engineering work covering upgrades, observability maintenance, architecture review and agency subcontracting. I do not provide on-call or 24-hour incident response.
I run a 3-node k3s cluster on mixed ARM and x86 hardware. Argo CD reconciles the workloads, Cilium handles networking, and Grafana Alloy collects metrics, logs and traces. Backup schedules, dashboards and most configuration live in Git. The manifests, playbooks and commit history are public.
g2net is a one-person consulting practice based in Lithuania. I work remotely with engineering teams across European hours, in English or Lithuanian.
I contract and invoice as a registered sole trader in Lithuania.
Per project when the scope is clear, or per hour when it is not. Terms are agreed in writing before work starts.
I take on limited project work alongside a full-time role. Ask for current timing and I will give you a realistic start date.
No. I do not take paging duty or provide 24-hour incident response.
I will say so during the first conversation. If I know a better route, I will point you to it.
A few sentences are enough: what you are running, what needs to change, and any constraints or timing I should know. I reply within two working days.
Enquiries are handled by direct email. No mailing list, analytics or CRM. Privacy