Skip to content
dabigc.com · Field notes of a platform engineer

I make
infrastructure
boring. On purpose.

Boring means the platform just works. Zero customer-impacting incidents since 2021.

Cole Conrad
COLE CONRAD
@dabigc

I'm Cole, a Principal Platform Engineer with 20+ years of building the infrastructure other people ship on. At HMH I co-architected a multi-cluster Kubernetes platform that 400+ engineers depend on every day, without losing sleep over it.

k8s + IaC + Cloud + AgentsPlatform Engineering
linkedin.com/in/dabigc github.com/dabigc

Ask me anything about my work.

grounded in a curated knowledge base
№ 01 · The approach

Layered, documented, and boring on purpose: that's how I build anything I expect to keep running. End-to-end ownership across design → implementation → operations → optimization, working close to the engineering teams who live on what I ship.

Build once, enable many. An internal developer platform should let a team go from idea to production by filling out a template. The measure of the work is how invisible it is.

"It runs. It actually runs. Now make it scale."the loop I've been in for two decades
№ 02

The work so far.

From the engineer clients asked for by name to the platform 400+ engineers lean on.

Principal DevOps Engineer

HMH Education · Remote
Mar 2019–Present
Senior → Staff → Principal

Platform engineer on a six-person team building internal developer platforms for 400+ engineers and tens of millions of daily end users. Promoted three times as scope grew.

  • Co-architected production Kubernetes across multiple clusters; cut time to first deploy from days to hours
  • Built an internal Claude Code plugin marketplace and codified agentic engineering workflows
  • Run observability as code: golden-signal monitors, SLOs, and error budgets for over a hundred services
  • Built an enterprise Terraform practice with reusable modules and TFE patterns; provisioning went from days to hours
  • Architected HashiCorp Vault at scale: dynamic and static credentials with least-privilege access across AWS and Concourse
  • Standardized ML platform onboarding: new tenants from a year to under two weeks
Case study № 01Building a Kubernetes Platform with Zero Customer-Impacting IncidentsEKS · multi-cluster · Vault · Consul · AmbassadorRead the case study →Case study № 02Architecting ML Infrastructure for AI-Powered Educationmulti-tenant ML Ops · SageMaker · Back-to-School deadlineRead the case study →Case study № 03CI/CD at Scale: Automating the Path from Code to ProductionConcourse catalog · GitHub Actions · 6,000+ repo migrationRead the case study →Case study № 04Codifying Agentic Engineering for a Platform TeamClaude Code marketplace · ship-pr · adoptable profilesRead the case study →
Featured case study № 05Running an Ops Platform as CodeOwnership, alerting, and staffing for hundreds of services, declared in code and validated in CI. Drift becomes a diff, and teams change their own on-call through a reviewed PR.Read the case study →
2% to 83%
ownership coverage in two weeks
Hundreds
of services cataloged and health-scored
Self-service
on-call changes by reviewed PR

Senior Systems Engineer

MXOtech · Chicago, IL
Oct 2014–Mar 2019
Sys Admin → Internal Eng → Senior

MSP supporting 100+ client environments. Grew from client-facing admin into the engineer trusted with the infrastructure underpinning the company itself.

  • Became the on-site engineer clients requested by name
  • Operated data centers under HIPAA, including Blue Cross Blue Shield of Illinois
  • Took a leading role in HITRUST certification, well beyond the role's scope
  • Introduced CI pipelines that moved devs off ship-straight-to-prod

Director of Information Technology

Gestalt Community Schools · Memphis, TN
Jul 2011–Jul 2014
Built the IT org from scratch

Charter network. Replicated a proven one-to-one technology model as the network grew from one school to five across four campuses, largely solo, on an aggressive timeline.

  • Scaled a 1:1 program from one school to five; imaged a thousand laptops and managed hundreds of iPads with Jamf
  • Delivered $1.5M in single-year savings by cutting overlapping digital curriculum
  • Unified access through Clever (then brand new) so one login reached content, the LMS, and PowerSchool
  • Replaced consumer gear with Cisco, Aerohive, and Barracuda networking for 2,000+ students and staff
№ 03

Education & Recognition

2004-2007

Stark State College of Technology

Multimedia Design and Development

Recognition
Standing Ovation AwardHMH · Q4 2022
Fred Factor AwardMXOtech · 2016
Eagle ScoutBoy Scouts of America
Conferences

KubeCon North America

2019, 2022, 2023, 2024, 2025 attendee

Connect
linkedin.com/in/dabigc github.com/dabigc
№ 04

The stack, honestly.

Tools I use daily. Depth over breadth.

Container Orchestration

Kubernetes (EKS)DockerMulti-clusterAutoscalingPlatform Engineering

Infrastructure-as-Code

TerraformTFEGitOpsConfiguration as CodeHCL

Cloud · AWS

AWSEKSEC2RDS/AuroraVPCIAMCloudFrontLambdaS3Route53

Secrets & Security

HashiCorp VaultConsulDynamic CredentialsAppRoleSOC2/GDPR

CI/CD & Automation

Concourse CIGitHub ActionsDeployment AutomationGitOps

Observability

DatadogGrafanaPrometheusInfluxDBMonitoring Architecture
№ 05 · The proof

This site is a running system, watched the way I watch production.

The chatbot above has retrieval, full tracing on every conversation, automated evaluations, and a private dashboard I read it through. The platform skills on this page also run the site.

69
automated tests across 9 eval datasets
7
layered guardrails on the chatbot
100%
of conversations traced in Langfuse
Daily
LLM-judge run

Let's talk.

I build platforms teams rely on to ship. If you need infrastructure at enterprise scale, I'd like to hear about it.

cole@coursecode.net linkedin.com/in/dabigc
Cole Conrad · Principal Platform EngineerPrivacy