RunForge — Cloud platforms and AI systems leaders can approve.

We work across EKS, GKE, and AKS. Our GitOps methodology is portable by design, but implementation details are adapted to each provider's IAM, networking, autoscaling, and managed Kubernetes behaviour.

Cluster configuration, security posture, GitOps maturity, cloud cost analysis, and architecture review, all in one engagement. You get a single detailed report with a prioritised remediation roadmap and commercial context for recommended next steps.

We manage the cluster (upgrades, node groups, networking, autoscaling) and platform components (ingress, cert-manager, monitoring). You manage your applications and deployment workflows. GitOps tooling (Argo CD) is included in the Premium tier, and available as an add-on for Essential and Professional.

RunForge is not SOC2 Type II or ISO 27001 certified. We operate with compliance-aware practices and can map Kubernetes posture against CIS Kubernetes benchmarks, Pod Security Standards, network policies, and secrets management expectations. For teams pursuing SOC2 or ISO 27001, we provide evidence-ready documentation, agreed control implementation support, and auditor-friendly architecture records.

Yes. Most work is remote. RunForge has Melbourne VIC + Driver NT, Australia business locations, which keeps collaboration straightforward across Australia, New Zealand, and APAC with overlap for engineering teams operating in local business hours.

All work is led by senior RunForge engineers. On-call coverage is provided by vetted team members briefed on your environment and held to the same bar before touching anything. No offshore. Work is never handed off without your knowledge.

Every engagement produces documented runbooks, architecture decision records, and operational playbooks, all committed to YOUR repository, not ours. Managed services clients have dedicated backup engineer coverage. Knowledge is never siloed in one person.

Each paid assessment SOW names the required readout artifacts, acceptance criteria, and remedy path before work starts, including the ranked risk map, prioritised roadmap, and fixed-scope next options.

Senior engineers do the work. Not juniors supervised by the person who sold the deal. Pricing is fixed per engagement, so the scope is clear before delivery starts. Everything we build comes with knowledge transfer — runbooks, architecture docs, and operational playbooks — so your team can run it without us. Our Melbourne VIC + Driver NT, Australia business locations make collaboration straightforward for Australian teams.

Week 1 is discovery. We get access to your clusters, codebases, CI pipelines, and existing documentation. Most engagements start with a Platform Assessment during this phase. Week 2, we start shipping. You'll see the first deliverables and a clear plan for what comes next.

You do. Custom code, configurations, and documentation built for you are yours once paid for. We retain ownership of our pre-existing tools and frameworks only. Full details are in section 5 of our terms of service.

RunForge is a new brand, so we don't have a long client list to point to yet. The engineers behind it have years of delivery experience across platform engineering, cloud infrastructure, and AI systems. We're happy to walk through relevant past work on the initial call. If you want a lower-risk starting point, begin with an assessment and use the output to decide what comes next.

For self-hosted AI work scoped in the SOW, Australian cloud regions are the default. Provider-hosted tooling, broader APAC deployment, telemetry, support access, backups, and hosted model APIs are documented as explicit exceptions before rollout. Some developer tools and workflow platforms are provider-hosted and may send telemetry or prompt data to their vendors. We help you evaluate those flows and keep exceptions explicit.

Only if you explicitly choose to. Our default recommendation is self-hosted inference using vLLM or Ollama on Australian infrastructure. If you want to use a hosted API, we help you assess the compliance implications and set up appropriate guardrails.

Codex, Claude Code, Cursor, and OpenCode. We handle SSO integration, policy configuration, agent permissions, repo skills and rules, evals, Browser QA, MCP server integration, PR handoff, and provider-hosted data-flow review. The goal is approved AI use inside your security and compliance framework.

Most scoped private-inference engagements are planned around 3 to 6 weeks from kickoff to a validated first deployment, depending on GPU availability, model selection, data-flow decisions, and acceptance criteria. The intended output is an OpenAI-compatible API endpoint in your infrastructure with monitoring, load testing evidence, and an operations runbook.

That's what AI Workflow Automation is for. We map your processes, design AI-augmented workflows, and deploy self-hosted n8n by default. If you prefer a provider-hosted workflow platform, we document that exception before implementation. Most engagements start from $5,000 and complete in 2 to 4 weeks.

The AI Workflow Automation engagement starts at $5,000. It's the fastest way to get a production AI workflow running. For a broader assessment, book a free 30-minute call and we'll scope what makes sense for your team.