02.1
Cloud architecture
From a single VPC to multi-region platforms. Account structure, networking, identity boundaries, data residency — designed for what the system has to do in three years, not what slides well today.
01 — Engineering practice · Est. 2017
Engineering for systems that have to work.
Vongoid is a senior engineering practice. We design, build, and stabilise the cloud platforms, distributed systems, and infrastructure that serious products depend on — then we hand them over, well‑documented, ready to run.
02
Capabilities
Six areas we work in. Each engagement is scoped around one of them, or the seam between two.
02.2
Platform engineering
Internal developer platforms that turn “how do I ship this?” into a single command. Service templates, paved roads, golden paths — without locking teams into a single opinionated stack.
02.3
CI / CD pipelines
Reproducible builds, deterministic deploys, and an audit trail that holds up under scrutiny. Pipelines that survive a 3am page, not just a demo.
02.4
Infrastructure modernisation
Untangling snowflake servers, hand-rolled deployments, and ten years of accumulated workarounds. We move systems off their original authors without freezing the product roadmap.
02.5
Reliability & observability
SLOs that mean something, dashboards an on-call engineer can actually act on, and a post-mortem culture that improves the system rather than assigning blame. We treat reliability as a product surface.
02.6
Technical leadership & advisory
A senior engineer or architect embedded with your team — for a quarter, a roadmap, or a hard problem. We help you make the calls, write the ADRs, and grow the people who will inherit the system.
03
Stance
Most platforms fail not because they were built wrong, but because they were built to be demonstrated, not operated.
We design for the second year of operation — the on‑call rotation, the security review, the day the original author is on parental leave.
-
i.
Production‑first thinking
Every component is sized, instrumented, and documented for the failure modes it will actually see — not the ones that look good in a diagram.
-
ii.
Fewer moving parts
We pick the smallest stack that solves the problem and resist the urge to add a new service every time a requirement shifts.
-
iii.
Operable, not just shippable
If a human can't triage it at 3am with a runbook and a dashboard, it isn't finished — it is half‑shipped.
-
iv.
Documented decisions, not just code
We leave behind architecture decision records, post‑mortems, and a written basis for every non‑obvious choice we make.
-
v.
Transferable knowledge
Our goal is that, six months after we leave, your team can change the system with confidence — and replace us without fear.
04
Method
How an engagement actually moves. Five phases, no surprises.
-
04.1
Diagnose
Read the code, read the runbooks, read the incident channel. We don't start with a proposal — we start with a shared map of the system as it actually is.
→ 1 to 2 weeks -
04.2
Design
Architecture, sequence diagrams, an explicit list of trade‑offs. We write it down so it can be argued with, not just admired.
→ ADRs · threat model · plan -
04.3
Build
Small, reviewable, instrumented. We work in your repositories, behind your code review, in front of your team. Nothing is shipped that your engineers can't defend.
→ merged in main -
04.4
Stabilise
SLOs, dashboards, runbooks, on‑call training, a real post‑mortem cadence. The system earns the right to be called production.
→ SLOs · on‑call · review -
04.5
Transfer
Written walkthroughs, recorded walkthroughs, and a deliberate handoff. We don't stay attached at the hip — we leave you more capable than we found you.
→ handover doc
05
Stack
Tools we are fluent in. Not a marketing matrix — a working inventory, grouped by what they actually do.
Cloud
- AWS — primary
- GCP — secondary
- Cloudflare — edge
- OVH · Hetzner — bare metal
Orchestration
- Kubernetes (EKS · GKE · self)
- ECS · Fargate
- Nomad
- systemd (for the ungovernable)
Infrastructure as code
- Terraform
- Pulumi
- Crossplane
- Ansible · bash
Languages
- Go
- Rust
- TypeScript · Node
- Python
Data
- PostgreSQL · Citus
- Kafka · NATS
- Redis · KeyDB
- ClickHouse · DuckDB
- S3 · object stores
CI / CD
- GitHub Actions
- Buildkite
- Argo CD · Flux
- GitLab CI
Observability
- OpenTelemetry
- Prometheus · Grafana
- Loki · Tempo
- Datadog (when required)
Security
- OPA · policy as code
- HashiCorp Vault
- SBOM · cosign · sigstore
- SOC 2 · ISO 27001 awareness
06
Contact
Ready to talk?
Whether you need a senior architect, a platform team, or an opinionated second pair of eyes — write to us. We respond within two business days, with a real engineer on the other end.
Or, if you prefer context first, send a one‑paragraph brief — the system, the team, the problem, the deadline.