Skip to content
Jack Devlin
B2B contracts · September 2026
// about me

Hey, I'm Jack

Independent B2B contractor based in the UK. I sit between data science and production, and I build the platforms that get models and services off a laptop and onto a Kubernetes cluster that doesn't fall over. Day job: Heimdall, a deployment dashboard 20+ engineers open every morning, a shared CI/CD library across 20 services, and the observability stack underneath. Night job: MSc in Artificial Intelligence, finishing August 2026.

Platform EngineeringMLOpsKubernetesAI InfrastructureObservability
scroll
<journey />

The Story

From graduate QA to platform engineering. Here's how I got here.

2023

QA Engineer

Graduated with a BSc in Computer Science. Jumped straight into a massive monolith → microservices migration, building the first CI/CD pipelines from scratch.

2024

Platform Engineer

Took ownership of the platform. Built observability from zero, implemented GitOps with ArgoCD, and designed deployment automation for 20+ services.

2025

Site Reliability Engineer

Standardised the pipeline platform onto one shared library across every service. Started the deployment-metrics tooling that became Heimdall.

2026

Platform & MLOps · MSc AI

Pivoting toward AI infrastructure. Finishing an MSc in Artificial Intelligence (August 2026). Operating as an independent B2B contractor, focused on the gap between data science and production.

<principles />

How I Work

A few things I've come to believe after a few years on platform teams.

01

Build for the morning standup

If a tool doesn't have a place to look, it doesn't get used. Most of my best work has been adding a UI to something that was already running headless.

02

Boring is a compliment

The ideal platform fades into the background. Engineers shouldn't have to think about it any more than they think about the office wifi.

03

Operability is a feature

If a teammate can't tell whether your service is healthy in under a minute, you haven't finished it yet. I default to one-curl health checks and a runbook.

04

Ship the diff, not the rewrite

Every project I'm proud of started as a small thing that quietly became load-bearing. Big-bang plans almost never survive contact with reality.

<stack />

Tech Stack

What I work with day to day.

Orchestration

KubernetesEKSArgoCDHelmKustomize

MLOps & AI

PyTorchMLflowKubeFlowNVIDIA GPU OperatorTriton

Observability

PrometheusGrafanaLokiTempoThanos

Cloud & IaC

AWSTerraformAWS CDKCloudFormation

Languages

PythonBashTypeScriptGo

Security

FalcoVeracodeIstioOPA

Data

KafkaPostgreSQLTimescaleDBRedis
<now />
⚙️

Right now

Shipping Heimdall · Finishing the MSc · Lining up September 2026

Wrapping up my current contract on the platform team I helped build. Heimdall is still the centre of gravity — a deployment intelligence dashboard used daily by 20+ engineers across 17 services.

Finishing an MSc in Artificial Intelligence in August 2026. The dissertation looks at deep learning applied to compute resource allocation — the same problem I keep running into on the platform side, so the two halves are converging.

Available for fully remote B2B contracts (Outside IR35 or international equivalent) starting September 2026. Platform engineering, SRE, and MLOps / AI infrastructure work.

Still reading?

Drop me a note. About a B2B engagement, an AI infrastructure problem, or anything that overlaps with the work above.