Embedded platform engineer
Monthly · rolling
One of us in your team, in your standups, on your board. Useful while you are hiring, or when a platform team of two needs a third pair of hands who has done it before.
Monthly rate · 30 days notice
Services
Six places engineering time disappears
We do not sell a named product. We work in your stack, on the specific thing that is slowing your team down, and we leave it in a state your engineers can maintain without us.
IDP
Internal developer platform
A paved route from empty repository to production, with the security, observability and deployment defaults already correct — so that the fastest path and the approved path are the same path. We start with the two or three journeys your engineers repeat most, not with a portal nobody asked for. Where the catalogue is the source of truth, we put an assistant on it, and expose the golden path to coding agents through MCP under the same permissions as humans.
- Service templates, scaffolding and golden paths
- Self-serve environments, databases and queues
- Backstage or equivalent, only where it earns its keep
- Catalogue-grounded assistant on the live service catalogue
- MCP so coding agents take the paved path under the same permissions
- Platform APIs so automation is not a copy-pasted script
CI / CD
Build and delivery systems
Pipeline archaeology first: where the minutes actually go, which jobs block which, and how often a green build is a lie. Then surgery — caching, dependency graphs, test sharding, remote execution and honest flake handling.
- CI rework across GitHub Actions, GitLab CI, Buildkite, Jenkins
- Build caching, sharding and remote execution
- Flake detection, quarantine and a policy for fixing them
- Progressive delivery, feature flags and rollback that works
IAC
Infrastructure and environments
Infrastructure as code that a product engineer can read without a translator, and environments that appear when they are asked for. Preview environments per pull request turn code review from imagination into inspection.
- Terraform, OpenTofu or Pulumi module design and refactoring
- Ephemeral preview environments per pull request
- Drift detection, policy as code and blast radius limits
- Cloud cost attribution that a team can act on
SRE
Reliability and on-call
Objectives tied to something a customer would actually notice, alerts cut down to the ones worth waking someone for, and runbooks written by the people who will be holding the pager at 3am. Then we practise, because an untested runbook is fiction.
- SLOs, error budgets and a policy for spending them
- Alert triage, deduplication and noise reduction
- Runbooks, incident roles and game days
- Blameless post-incident review your team will keep doing
DXR
Developer experience research
Measure before you build. We combine pipeline telemetry with structured interviews and a survey your engineers will finish, then give you a baseline specific enough to argue with. Most clients start here, and some stop here quite happily.
- DORA and SPACE metrics from your own systems
- Developer interviews and friction logging
- Ranked, costed list of what to fix first
- A repeatable measurement you can re-run without us
ENABLE
Platform team enablement
A platform team is a product team whose users happen to sit nearby. We help yours behave like one: a roadmap, real users, adoption measured rather than assumed, and the nerve to deprecate things properly.
- Operating model, team topology and interfaces
- Platform roadmap, intake and prioritisation
- Adoption metrics and deprecation practice
- Coaching for a first-time platform lead
Engagement models
Four shapes, all of them written down first
Most work is embedded in your team. The other shapes are for a scoped build or a short measurement phase. Whichever you pick, the scope is agreed before work starts and changes go through the client desk with a costed impact assessment.
Platform build
4–12 weeks · fixed scope
A named set of deliverables with acceptance criteria, built embedded with your team, in your repositories, demoed weekly. Handover and a 30-day review are part of the price.
Priced per scope after the baseline
Fractional platform lead
1–2 days a week
Direction without a full-time hire: roadmap, technical decisions, vendor calls and coaching for the engineer who will eventually take the role permanently.
Minimum three months
Baseline audit
2 weeks · fixed fee
We measure where the time goes and hand you a written report with a ranked, costed list of fixes. No obligation to do any of them with us, and the report is yours either way.
Fixed fee, quoted up front
Tooling
We work in what you already have
Migrations are a cost, not an achievement. If your stack is boring and it works, we will leave it alone and fix the thing that is actually hurting.
- AWS
- GCP
- Azure
- Kubernetes
- ECS
- Terraform
- OpenTofu
- Pulumi
- Helm
- Argo CD
- GitHub Actions
- GitLab CI
- Buildkite
- Jenkins
- Backstage
- AWS Bedrock
- MCP
- Nix
- Bazel
- Docker
- Prometheus
- Grafana
- OpenTelemetry
- Datadog
- PagerDuty
- Vault
- OPA
Not on the list? Ask. The principles travel further than the logos.
Start with the baseline
Two weeks, a fixed fee, and a report you can take to your board — whether or not the next stage is with us.