How we work
We work alongside client teams on cloud platforms that carry revenue: landing zones, Kubernetes platforms, and CI/CD paths. We aim for deliverables they can extend once we step back.
Internally we shorthand that mindset as the Engineering Circle: change ships with bounded blast radius; improvements show up in measurable reliability; learning happens when telemetry contradicts an assumption; sharing means artifacts someone else can rerun, not slides.
Specs before improvisation
RFCs, ADRs, and runnable docs beat hallway lore. When IAM boundaries, network paths, or blast radius are still fuzzy, we tighten them before merge, not during an incident call.
Changes earn evidence
Benchmarks, failure-mode reviews, and sober pager assumptions temper roadmap optimism. Safer batches through pipelines beat novelty on slide five.
Interfaces at work
A flat org chart isn't fuzzy accountability: what you own, who decides, how pairing fits the week, where to escalate. It should be explicit. Pushback on risk or scope is normal when it's concrete.
Incidents teach the system
Reviews stress sequencing and controls, the same instinct as a blunt postmortem, not hero narratives. That carries into incident threads, careful code review, and treating customer engineers as peers in the same outage channel.
Knowledge ships as artifacts
Demos and pairing when they shorten feedback loops; otherwise defaults land in repos and modules your teams own. Cleverness that lives only in someone's head doesn't survive rotation. We treat undocumented wins as unfinished.
What we build
We build cloud platforms and the integrations around them: typically alongside a customer’s platform, SRE, or cloud centre of excellence. That means AWS or Azure landing zones, identity and access, security guardrails, CI/CD, service catalog patterns, and glue between systems so product teams can ship with less friction. Most of it is expressed as code; we standardise on Terraform where it fits the problem.
Our work lands with in-house development teams: repeatable environments, clear handover, and controls that match business and security policy. Day to day, that is a lot of structured problem-solving within organizational and technical boundaries: finding the smallest change that makes the next safe step obvious.
Example. We needed Entra ID (Azure AD) group membership to drive temporary permissions in AWS faster than the default directory sync cycle allowed for a programme. Rather than accepting a 40-minute lag, we designed a small, auditable automation path to trigger sync on a tighter cadence: so developers could get short-lived access without weakening access-review trails.
Across clients, the habits stay the same: automate what repeats, store intent in Git, deploy through pipelines, and review before production. That is both how we reduce toil and how we make sure teammates can improve what we leave behind.
How the bar stays high
Senior engineers don’t stay senior on their own. These are the practices we run on purpose, so depth compounds instead of decays. They are also, frankly, what makes the work interesting.
WTF sessions
We run What The Faqtory! (WTF) on a rhythm we aim to keep monthly: engineers walking through real customer work, technical deep-dives, demos, home experiments, and what didn’t work. Tribal knowledge ends up in the room instead of in three people’s heads.
Time and tools to learn
Pluralsight and A Cloud Guru access, a real budget for certifications, and study agreements where the topic justifies it. The cloud toolchain doesn’t stand still; we expect the team to move with it, and we pay for the time it takes.
In the community
We attend and speak at meetups, follow the AWS, Azure, and Kubernetes communities, and back engineering events that keep the field honest. Signal lives where engineers compare notes, before it becomes a vendor slide.
Numerous internal sessions throughout the year; we aim for a monthly rhythm.
The mix changes with what the team is shipping. Topics we keep coming back to:
- Landing-zone patterns at scale
- Karpenter migrations
- Cilium for cluster networking
- Service-mesh trade-offs in production
- GitOps with Argo CD vs. Flux
- Cost-aware autoscaling on EKS
- RAG inference: the boring infrastructure choices
- Vault disaster-recovery rehearsals
- Multi-account AWS governance
- FinOps tagging across teams
- Observability: three real configurations
- Postmortems people actually read
Tools and services we use
A non-exhaustive list of the technologies, tools, and services we work with day to day, including stacks behind AI inference and integrations.
Anthropic
OpenAI
Kubernetes
Docker
Traefik
HashiCorp
Packer
Terraform
Vault
Elastic Stack
Splunk
Datadog
GitLab
GitHub
Jenkins
Serverless
Node.js
NGINX
Apache
Ansible
Puppet
Chef
Python
Ruby
PowerShell
Bash
SonarQube
MySQL
SQL Server
MongoDB
Confluence
