James Thurman

9.0

@jmichaelthurman

Platform Engineer | AWS & Kubernetes at Scale | AI Governance

Oklahoma City Metro

AI Fluency Score

9.0/10

Assessed 3/31/2026

Velocity

HIGH

About

I build the infrastructure layer that engineering teams stand on.

Currently part of the platform team at SumerSports, where we operate infrastructure for 100 engineers across 10 EKS clusters and 15 AWS accounts. I designed a self-service provisioning framework that cut developer wait times from 10 business days to 30 minutes, and I maintain eks-upgrade-check — an open-source CLI (v3.4.0, 123+ PRs) that scores Kubernetes upgrade readiness and runs in production as a weekly automated scan across our clusters.

Before that, three and a half years as an SRE at Moonswitch across twelve Kubernetes clusters, five clients, and compliance frameworks from FedRAMP to PCI-DSS. At Outcomes Health, I ran a 12-month program that reclaimed $600K+ from an undocumented $1.2M Azure footprint, with CTO sponsorship and executive reporting.

I also design AI governance systems. My claude-config framework — controls mapped to SOC 2, NIST AI RMF, and OWASP LLM Top 10 — is piloting among the platform team at SumerSports. Not a thought experiment; a working evaluation and enforcement system.

AI Wingman Summary

AI Generated

Generated 3/31/2026

James Thurman is a Platform Engineer with deep expertise in AWS, Kubernetes, and AI governance, building the infrastructure layer that engineering teams depend on at scale. An expert-level AI practitioner, James designs production-grade agentic workflows that deliver measurable results — including a slash-command pipeline that reduced multi-week IaC bundle authoring to a single automated sequence, generating a consistent 4-5x productivity multiplier. His meta-prompting approach is genuinely innovative: using Claude Opus to author prompts optimized for reliable execution by downstream models, essentially engineering AI-to-AI orchestration. He maintains a 56-file, 31,000-word methodology framework with bootstrap portability across repositories, and has built AI governance controls mapped to SOC 2, NIST AI RMF, and OWASP LLM Top 10 — currently in production pilot. James brings a "code as product, not craft" philosophy to every engagement, transforming experimental AI tooling into documented, scalable systems that teams can build on.

Work Experience

DevOps (Platform) Engineer at SumerSports

Latest

SumerSports is an AI-first sports analytics company. I'm part of a distributed platform team spanning the US, Serbia, and Kenya, operating infrastructure for 100 engineers across 10 EKS clusters and 15 AWS accounts. My title is DevOps Engineer; the scope is broader — I own the self-service provisioning framework, architecture RFCs, upgrade readiness automation, and on-call rotation.

Key accomplishments:

Designed the Terramate Catalyst Bundles framework, cutting developer infrastructure provisioning from 10 business days to 30 minutes across Aurora Global clusters, CloudFront, SQS, SNS, S3, and more
Filed 70+ infrastructure drift remediation issues across eight environments
Authored strategic architecture RFCs for blue-green EKS cluster upgrades (v1.30→v1.33) and Transit Gateway monolith decomposition
Integrated eks-upgrade-check — my open-source CLI (v3.4.0, 123+ PRs) — as a containerized weekly scan across all clusters, automatically opening GitHub issues for remediation
Carry platform on-call (Datadog): 34 tickets, 88% resolution rate

Senior SRE at Moonswitch (Platform Consulting)

Moonswitch is a managed services provider. I operated twelve EKS/GKE clusters and two air-gapped Azure Gov Cloud environments across five clients, each with different compliance requirements — FedRAMP, FISMA, NYDFS, SOC 2, and PCI-DSS. I served as technical account manager and carried 24/7 on-call for Gold-tier production clusters.

Key accomplishments:

Rebuilt a client's disaster recovery baseline from 72 hours to 12–18 hours (75%+ reduction) by implementing Terraform IaC with drift detection across AWS, Azure, and GCP — after a malicious failure event exposed that their documented DR process had never been tested
Diagnosed critical instability in a 65-node EKS cluster running 250+ preview environments — root causes were three compounding failures: DNS quota exhaustion, GraphQL middleware generating 800+ concurrent database connections, and application-level PSQL query issues. Built the observability stack (Datadog, New Relic, Elastic, Prometheus) that made all three visible
Operated across AWS, Azure, and GCP simultaneously, including FedRAMP air-gapped RHEL configuration for CUI workloads in a defense-sector client environment
Ran code reviews across NodeJS, TypeScript, C#, and Ruby as part of staff augmentation into client engineering teams

Links

Website

jmichaelthurman.com

Github

github.com/jmichaelthurman

Get Your AI Fluency Score

Join James Thurman and thousands of professionals measuring their AI skills. Get assessed, showcase your score, and improve your AI fluency.