Graham Clark

Platform engineering leader: cloud, Kubernetes and identity

Profile

Platform engineering leader with 20+ years in infrastructure, the last decade leading DevOps and platform teams on AWS, Azure and Kubernetes. Head of CIAM Platform at Alter Domus, leading a team of 5 running an Auth0 platform for ~55,000 users, and previously built and led its Platform Engineering team. Earlier, built DevOps teams from scratch at getaroom.com and People's Postcode Lottery. Early adopter of AI-assisted engineering: helped pilot Claude Code (under 100 users) ahead of a company-wide tech rollout.

Work experience

Alter Domus

Jul 2022 – Present

Head of CIAM Platform

Nov 2025 – Present

Own the Customer Identity and Access Management (CIAM) platform strategy and delivery, leading a team of 5, three of them contractors. The CIAM specialism is one of several T-shaped sub-teams Platform Engineering was restructured into, each splitting time 50/50 between specialism work and shared wider-platform duties. The CIAM platform is in SOC 2 scope for several applications.

  • Designed and launched an Auth0-based CIAM platform serving around 55,000 users, comprising multiple Kubernetes-based apps (including a custom MFA adapter, support portal, and golden path sample apps) and supporting AWS infrastructure (EventBridge, CloudWatch, Connect for calls, SNS for SMS)
  • Led the migration of in-house applications from the legacy B2C platform to the new Auth0-based platform, cutting auth-related support tickets by 50%
  • Built self-service Claude access into Alter Domus's central platform portal: users request model access and usage limits, managers approve and see their own budget
  • Advised Alter Domus's AI engineering team on identity architecture for their AI assistant and its MCP tool servers, delivering service-to-service and agent authentication (Auth0 client-credentials flows, MCP gateway auth), now in production
  • Evaluated Kong AI Gateway and Envoy AI Gateway against LiteLLM as the company's LLM gateway; recommended LiteLLM, avoiding additional cost and vendor lock-in
  • Adopted Claude Code for AI-assisted infrastructure-as-code and application development across the CIAM platform's services

Auth0, Azure, AWS, Kubernetes (EKS), GitHub, LiteLLM, MCP (Model Context Protocol), Claude Code

Senior Manager, Platform Engineering

Jul 2022 – Nov 2025

Led a Platform Engineering team of 3 engineers, delivering cloud infrastructure across Azure and AWS.

  • Built Data and Analytics' first cloud platform on Azure, delivering a capability that hadn't previously existed for the function
  • Established AWS foundational platforms, including an EKS-based Kubernetes golden path
  • Implemented Okta to integrate an acquired company's application estate, meeting their existing IdP requirements as part of the M&A onboarding
  • Part of the founding group that piloted AI adoption at Alter Domus (initial pilot of under 100 users), building reusable Claude Code skills, MCP servers, GitHub PR tools, customised Claude Code installers, golden paths and an onboarding and user-management model; later rolled out to the entire tech team

Azure, AWS, Kubernetes (EKS), Terraform, GitHub, Okta, Claude Code, MCP (Model Context Protocol)

Trunomi Startup, since closed

Aug 2021 – Jul 2022

Principal DevOps Engineer

Responsible for the Trunomi platform and supporting infrastructure, reporting to the CEO and CTO.

  • Delivered a production-ready Kubernetes environment solo, in 10 working days in month two at the company, against an external firm's quote of 44 person-days for the same scope
  • Implemented robust monitoring and alerting solutions for Trunomi's production environment
  • Documented and automated the deployment of the Trunomi platform

AWS, Kubernetes (EKS), Kyverno, Terraform, ArgoCD, CircleCI, GitHub Actions, Datadog, Python, Go, Rust

getaroom.com

Mar 2019 – Aug 2021

DevOps & Platform Operations Manager

Jan 2021 – Aug 2021

Led the DevOps and Platform Operations teams (4 engineers each) within a 25–30 person engineering organisation, defined cloud strategy and reported at board level.

  • Guided the Platform Operations team towards an SRE focused methodology
  • Worked with the Revenue and Affiliate teams to identify key SLIs allowing us to reduce load on the platform by ~30%
  • Achieved a cultural shift within the company applying DevOps principles across the development teams

Lead DevOps Engineer

Mar 2019 – Jan 2021

Grew the internal DevOps team from a one-person function to 4, responsible for the stability, reliability and security of the getaroom platform.

  • Spearheaded a migration of the getaroom platform to Kubernetes
  • Reduced release/deployment times from over 6 hours to less than 30 minutes
  • Reduced AWS spend by over 50% (~$50k per month) in early 2020 while ensuring we had a plan to allow for rapid growth as the effects of the pandemic subsided
  • Cut monitoring costs by ~25% by analysing our Datadog and New Relic usage, and logging costs by up to ~80% working with the development teams

AWS, Kubernetes (EKS), Terraform, ArgoCD, Jenkins, GitHub Actions, New Relic, Datadog, Python, Go

ECS Digital DevOps consultancy

Sep 2018 – Mar 2019

Senior DevOps Consultant

  • Delivered DevOps transformation consulting for client organisations, including RBS's 'Sendmoney' project pilot for Natwest
  • Recruited, mentored and trained other members of the ECS Digital team

People's Postcode Lottery

Aug 2016 – Sep 2018

Lead DevOps Engineer

Mar 2017 – Sep 2018

  • Managed the PPL DevOps Team and led the migration of platforms and services into AWS
  • Built a DevOps team from scratch (joining as the sole DevOps engineer, growing it to 5), delivering the 'Sqills Transformation Project' (the biggest project undertaken by the PPL development team to date) within an extremely tight timeframe
  • Led the platform delivery for 'Miljoenenjacht', a TV show run by the Dutch Postcode Lottery, and mentored other Novamedia lotteries in DevOps practices

DevOps Engineer

Aug 2016 – Mar 2017

  • Formulated and implemented PPL's DevOps strategy, owning its private cloud and customer-facing systems, and mentored the development and IT Ops teams in DevOps and Linux practices

Earlier Career

2000 – 2016

Systems administration and engineering roles at Red61, The Scottish National Blood Transfusion Service, Intrallect Ltd, George Watson's College, Scottish Cultural Resource Access Network, and The Medical Research Council - Human Genetics Unit, followed by Operations Manager at No Need 4 Mirrors, leading the support team and redesigning its customer-facing infrastructure with a resilient virtualised architecture.

Personal projects

TomewardSelf-Hosted Ebook Library & Reader

  • Built a Calibre/Kavita-style ebook library and reader (Rust, SvelteKit, SQLite) with an LLM-generated "story so far" recap for partially-read books, via a provider-agnostic OpenAI-compatible client supporting cloud APIs and self-hosted models (vLLM, llama.cpp, Ollama, LM Studio)

Homelab GitOpsFlux-managed Kubernetes

  • Run self-hosted infrastructure from a Flux GitOps repo, including a LiteLLM proxy routing to local vLLM, llama.cpp and Ollama backends over Tailscale, later extended to OpenRouter