Jens Skott
GitHubLinkedIn

Jens Skott

Drop-in platform engineer for GCP-native teams. I build the foundation — landing zone, IaC, observability, security and cost — and hand it over documented.

Platform and engineering leader who still ships. I have run a 20-person function on a £15M budget, and I am the sole platform engineer behind a product with 60,000 paying subscribers.

25 years in infrastructure and operations · London

I'm hiring for
A contractA permanent role
Basis
Outside IR35 preferred; inside considered
Where
Remote, or London on site
Seeking
Head of Platform / Principal Engineer
Where
London, or remote UK
Email, phone and day rateEmail, phone and package

Professional Summary

I take GCP-native companies from engineers deploying by hand to a real platform — landing zone, infrastructure as code, observability, zero standing access and cost control — and I have done it twice: across a 20-person engineering function at Checkatrade, and single-handed at Scoreline, where 60,000 paying subscribers run on what I built. Twenty-five years in infrastructure and operations, the last decade on GCP and AWS.

Key Highlights

  • ▸Sole platform engineer for a revenue-bearing consumer SaaS on GCP: Pulumi landing zone, factory-pattern service components, 45 architecture decision records, and about 460 merged pull requests in six months
  • ▸Reduced infrastructure incident hours by 75% and CI/CD runtime by 60% in leadership and consulting roles
  • ▸Zero standing access on GCP: Workload Identity Federation, Privileged Access Manager entitlements, no service-account keys anywhere I run
  • ▸OpenTelemetry-first observability with a vendor-swappable backend, plus request-based CPU billing defaults, sandbox spend guards and billing anomaly alerting; cut a retired product line from $3,105 to about $400 a month
  • ▸Built a read-only AI incident investigator that reads logs, traces, source and billing data and posts a structured assessment to Slack, on a provider-agnostic LLM port — and Jeeves, the team's Slack assistant ("Ask Jeeves") that answers platform questions from the documentation
  • ▸Ships product end to end solo — three iOS apps and one Android app on the stores, Swift 6 and Kotlin, with engine parity enforced by a golden-vector harness
  • ▸Terraform, Pulumi, GitOps (Spacelift, ArgoCD), Cloud Run, Cloud Functions, Cloudflare Workers, BigQuery + Dataform

Professional Experience

Senior Platform Engineer

February 2026 – Present

Scoreline (Fantasy Football Hub)

  • ▸Platform Foundation:Sole platform engineer for a subscription sports-analytics business on Google Cloud. Built the organisation landing zone and delivery platform in Pulumi (TypeScript, state in Cloud Storage): factory components that give every Cloud Run service and job observability, cost guards, IAM and secrets by default, ephemeral sandbox stacks per pull request, and a GitOps/ChatOps operating model — recorded in 45 architecture decision records. Between April and September alone: roughly 440 platform tickets closed and 460 pull requests merged across the two platform repositories. Left it handed over: a successor roadmap, an estate liveness register, and continuation guides for each bespoke system.
  • ▸Observability, AI Incident Investigation & Jeeves:Standardised on OpenTelemetry as the only telemetry surface in application code, with the backend behind a collector sidecar so it stays swappable, plus trace-correlated logging and job monitors keyed on absence of success. Designed and built a read-only AI incident investigator that picks up Datadog alerts from the alert bus, investigates against logs, metrics, traces, source and billing data, and posts a structured root-cause assessment to Slack — on a provider-agnostic LLM port. Built Jeeves, the engineering team's Slack assistant ("Ask Jeeves"): it answers platform questions from the indexed documentation with cited sources, carries every alert and investigation card with acknowledge, resolve and create-ticket actions, reports idle sandbox environments and their cost, and posts a weekly pull-request digest.
  • ▸Security, Access & Resilience:Moved the estate to zero standing access: two-tier Workload Identity Federation for CI, Privileged Access Manager entitlements for humans, least-privilege service identities, org policy and Security Command Center, and secrets consolidated out of CI variables into Doppler. Built the disaster-recovery backup service — Cloud Run Jobs backing up Postgres, Cloud SQL and object storage from a single typed manifest into an isolated backup project, with freshness monitors — and the runbooks behind it.
  • ▸Cost:Owned cost visibility and efficiency. Made request-based CPU billing the platform default and moved six production services onto it; environment-derived defaults cut the main API's staging compute cost by 80%; and wound a retired product line down from 35% of total GCP spend ($3,105 a month) to about $400. Sandbox spend guards are enforced in the factories, and billing anomalies route into the investigator.
  • ▸Subscription Platform:Re-platformed the tier-1, revenue-bearing subscription service that normalises Stripe, PayPal and RevenueCat webhooks into one event-sourced store — Bun and Hono on Cloud Run, Drizzle on Neon Postgres, CI-owned migrations, an admin operations console, and OpenTelemetry tracing end to end.

Founder & Principal Engineer

2015 – Present

Thoughtgears

  • ▸Infrastructure & standards:All GCP, Firebase, Cloudflare, GitHub and Workspace infrastructure in Terraform via Spacelift, zero-drift policy, no service-account keys anywhere. An org-level standards repo of decision records every product repo inherits.
  • ▸Nodrik — multi-tenant SaaS on GCP:Designed and built a read-only multi-tenant service that runs inside a customer's own Google Cloud organisation: Cloud Run scale-to-zero, Firestore, Pub/Sub, a Vertex AI Gemini tool-calling agent behind a swappable provider port, Stripe self-serve billing, one service account per tenant, a hard per-request spend ceiling, Terraform owning the Cloud Run services, CI through Workload Identity Federation.
  • ▸Analytics platform:A BigQuery warehouse with Dataform (raw → staging → marts, 100+ models across eleven brands) fed by TypeScript Cloud Functions collectors for Cloudflare, App Store Connect, Apple Search Ads, Google Play and Stripe — scheduled, private, secrets from Secret Manager, deployed through Workload Identity Federation.
  • ▸Web applications:React single-page apps on Vite and TanStack Router with a shared Radix component library; sixteen Astro sites on Cloudflare Workers Static Assets, including a paid product with Stripe checkout, magic-link auth and state in Cloudflare's edge key-value store.
  • ▸Consulting:Cut deployment times by 75% (two hours to thirty minutes) with Docker Compose and Firebase emulator workflows, moving a team from 3–4 releases a week to daily. For a major insurance client, cut CI/CD runtime by 60% and enabled on-demand releases.
  • ▸Also ships product end to end, solo:Three iOS apps and one Android app on the stores — Swift 6 and SwiftUI, SwiftData with CloudKit sync and versioned migrations, StoreKit 2, watchOS; Kotlin and Compose on Android — with a golden-vector harness enforcing engine parity across both platforms.

Infrastructure Team Lead

August 2025 – February 2026

Fresha

  • ▸From interrupts to a request model:Replaced ad-hoc, interrupt-driven support with a chat-based intake feeding a Kanban board, so work arrived as tracked requests rather than shoulder-taps. Shifted the team's balance from reactive firefighting to planned platform work, cutting context switching and lifting morale.
  • ▸Self-service over shoulder-taps:Raised documentation standards across the team's surface area so engineers resolved the common requests themselves, deflecting repeat questions away from the infrastructure team and shortening the path from problem to fix.
  • ▸Platform strategy:Proposed a three-year platform roadmap centred on developer experience — a 'Heroku-lite' abstraction over the tooling landscape.

Head of Engineering & Principal Engineer

February 2021 – March 2025

Checkatrade

  • ▸Operational Excellence:Reduced infrastructure-related incident hours by 75% (40h/month to 10h/month) by architecting a new event-driven platform on GCP as part of the AWS → GCP migration, and implementing rigorous Correction of Error frameworks.
  • ▸Leadership & Strategy:Managed a £15M annual budget and a headcount of 20, executing a 5-year strategic reorganisation that shifted the engineering culture from reactive operations to continuous innovation; introduced GitOps with Terraform and ArgoCD, cutting deployment times by a further 60%.

DevOps Lead Engineer

April 2020 – January 2021

Loveholidays

  • ▸Incident Response:Engineered ChatOps tools that reduced Mean Time To Acknowledge (MTTA) by 50% and Mean Time To Recovery (MTTR) by 10%, and mentored teams onto Terraform and Kubernetes for automated deployment.

DevOps Manager

February 2018 – April 2020

Deltatre OTT

  • ▸High-Performance Engineering:Personally developed a Golang API router capable of scaling from 5k req/sec to spikes of 1 million req/sec using horizontal pod scaling and traffic stickiness; led the subsequent migration to AWS EKS.

DevOps Engineer

June 2016 – February 2018

Accedo TV

  • ▸Cloud Infrastructure & Backend:Designed and ran AWS infrastructure in Terraform, Chef and ECS for a move to microservices, contributing Go backend code to the decomposition of the monolith.

DevOps Engineer & IT Manager (early career)

2009 – 2016

Schibsted · Klarna · King · Gears of Leo

  • ▸Early platform and container work:Built an Internal Developer Platform for VMware at Schibsted in 2011, ahead of such platforms becoming industry standard; OpenStack/Chef automation for King, an OpenShift dev environment at Klarna, and an automated Rackspace/Chef/Jenkins environment at Gears of Leo with early Docker and Kubernetes evaluation.
  • ▸Early 2000s – 2009:Network and systems administration across a series of small companies in Sweden — the hands-on infrastructure grounding underneath everything since.

Recent work

  • Platform foundation for a subscription business

    Sole platform engineer on a consumer SaaS: organisation landing zone in Pulumi, factory components giving every service observability, cost guards, IAM and secrets by default, ephemeral per-PR environments, 45 decision records, and a documented handover.

  • Operating model for an infrastructure team

    Replaced interrupt-driven support with a chat-based intake on a Kanban board, and raised documentation standards so engineers resolved the common cases themselves — reactive firefighting to planned platform work.

  • Security, cost and release engineering in insurance

    Zero standing access: Workload Identity Federation for CI, entitlements for humans, no service-account keys anywhere. Cut CI/CD runtime by 60% and moved the team to on-demand releases.

What I am looking for

  • A scale-up on GCP

    Somewhere platform is a real function with a mandate, not a rota bolted onto a product team.

  • A team to build

    I have hired, mentored and reorganised one before — 20 engineers at Checkatrade, and a delivery process at Fresha that moved the team off interrupts.

  • Hands on the keyboard

    I lead and I still write the Terraform. A role that takes me away from the work entirely is the wrong one.

Selected Work

  • Nodrik

    AI incident investigation for GCP-native teams

  • SteelRep·LastLift·Vargr

    Native strength & hybrid-training apps, iOS + Android

  • Passline

    UK fitness-test prep — Cloudflare Workers and Stripe

Core Competencies

Cloud & Platforms

GCPAWSCloud RunPub/SubBigQueryFirebaseFirestorePostgreSQLCloudflare WorkersWorkload Identity Federation

Infrastructure & Delivery

TerraformPulumiSpaceliftGitHub ActionsGitOpsKubernetesLinux

Observability & Operations

OpenTelemetryDatadogPrometheusGrafanaIncident Management / SREFinOps / Cost Optimisation

Languages & Frameworks

TypeScriptGoPythonSwift 6 / SwiftUIKotlin / Jetpack ComposeNode.jsBunSQL

Web & Frontend

ReactAstroTailwind CSSViteTanStack RouterDesign Systems

Mobile & Product

SwiftData + CloudKitStoreKit 2watchOSPlay BillingStripe

AI & Data

LLM Tool-Calling AgentsVertex AI (Gemini)Claude CodeDataform

Leadership

Team BuildingMentoring