NNakama Labs
Platform Engineering, written down

Defining the modern Platform Engineering standard for startups.

Define your platform before it defines you.

Nakama Labs helps startups build manageable engineering platforms from day zero through practical standards, playbooks, technical writing, and advisory work.

platform/decisions.log
$ nakama log --decisions
D-001 source control → monorepo + trunk
D-014 infra → Terraform, env per dir
D-022 cost → budgets + labels day 1
D-031 observ. → SLOs before alerts
WARN 3 decisions still undocumented
$
// The belief

Most startups do not fail at infrastructure because their engineers are bad.

They fail because platform decisions are made too late. The chaos is invisible — until it is not.

01CI/CD gets slower, one merge at a time.velocity
02Terraform becomes risky to run.change
03IAM becomes unclear — who can do what?security
04Cloud costs become invisible.cost
05Observability arrives after the incident.reliability
06DevOps is hired after the mess already exists.too late
// The Nakama Platform Standard

An evolving, public framework for startup platforms.

Ten layers every growing engineering team has to get right eventually. The Standard says what good enough for now looks like at each one — and what to revisit next.

P-01

Source Control & Monorepos

One clear place code lives, with branching everyone understands. The repo shape is a platform decision, not a personal preference.

What good looks like
Trunk-based with short-lived branches
Ownership encoded in CODEOWNERS
Repo strategy chosen on purpose, not by accident
// Advisory

Nakama Platform Blueprint

A focused advisory engagement for startups that need clarity before scaling their engineering platform. One pass over the whole platform, then a roadmap you can act on.

scope~3–4 weeks, async-first
outputa 90-day roadmap + decision records
Request a Blueprint →
01Platform maturity assessment
02Architecture & infrastructure review
03CI/CD & release process review
04Terraform / IaC structure review
05Cloud cost & billing review
06Security & IAM review
07Observability & incident readiness review
0890-day platform roadmap
09Practical playbooks & decision records
// Labs

Built, tested, documented.

Nakama Labs publishes experiments, not generic tutorials — real setups, run end to end, with the trade-offs written down.

EXP-01Running MCP servers on KubernetesA lab note on packaging and operating MCP servers inside Kubernetes.In progress
EXP-02Evaluating AI agent frameworksA comparison framework for testing agent tooling against real platform tasks.In progress
EXP-03Comparing observability stacksA small-team comparison of observability stack tradeoffs, cost, and operational weight.In progress
EXP-04Building an internal RAG systemA practical experiment for making internal engineering knowledge searchable.In progress
// About

Turning hard-earned infrastructure lessons into standards other teams can follow.

Nakama Labs is led by Ramiro Castillo, a DevOps and Platform Engineer with hands-on experience across cloud infrastructure, Kubernetes, GKE, Terraform, GitOps, Agones, Cloudflare, observability, CI/CD, and production systems.

The work is not about making infrastructure sound complicated. It is about turning the lessons learned in production into practical standards other teams can read, challenge, and use before the messy parts become expensive.

~/ramiro/stack.txt
KubernetesGKETerraformGitOpsArgo CDAgonesCloudflarePrometheusGrafanaCI/CDGCP

Have a platform decision you are unsure about?

If your startup is growing and your infrastructure is becoming harder to understand, let’s talk.

Contact Nakama Labs →
hello@nakamalabs.io