AI DevSecOps Engineer
I'm Warren McDonald. I harden Kubernetes, codify infrastructure with Terraform, and build defense-in-depth guardrails for generative AI systems.
A decade of DevOps, now focused on making AI workloads safe to run in production.
AI DevSecOps Engineer specializing in securing generative AI systems, orchestrating resilient Kubernetes infrastructure, and automating cloud-native DevSecOps pipelines. Proven expertise in hardening Google Kubernetes Engine (GKE) and AWS EKS environments, enforcing Zero-Trust workload isolation, implementing defense-in-depth LLM guardrails (OWASP Top 10 for LLMs), and codifying immutable infrastructure with Terraform.
The page you're reading runs in a hardened, non-root container on GKE, provisioned with Terraform and shipped through a signed, scanned GitHub Actions pipeline. The AI assistant below sits behind rate limiting, prompt-injection filtering, canary tokens and egress DLP, and you're welcome to try to break it.
Try the red-team sandboxWhere I've built and secured platforms
Twelve years across healthcare, fintech, media and telecom.
Led comprehensive DevOps transformation, orchestrating cloud infrastructure, Dockerizing applications, and implementing modern CI/CD pipelines in GitHub Actions. Architected and maintained highly available AWS Kubernetes clusters. Drove AWS to Azure migration with Terraform and led the implementation of Databricks in Azure for healthcare AI/ML workloads.
Provisioned and managed AWS infrastructure with Terraform, Kubernetes, Fargate, and EC2. Transitioned DNS to Cloudflare, configured Pritunl VPN with VPC peering, and deployed cross-region RDS read-replicas with Prometheus/Grafana monitoring.
Managed enterprise AWS deployments using Chef/OpsWorks and Jenkins. Built hardened AMIs with Packer and Docker, implemented autoscaling policies reducing timeouts by 90% during traffic surges, and managed offshore engineering teams.
Streamlined CI/CD with CircleCI, deployed Node.js/PostgreSQL services with AWS CodeDeploy, optimized high-traffic media properties with Fastly VCL CDN, and maintained centralized ELK logging stacks.
Maintained AWS environments using Terraform, built AMIs via Bamboo and Ansible, and deployed open-source observability solutions (Graylog, Netflix OSS).
Deployed Java services in Tomcat/Jetty with Splunk logging, executed Puppet Blue/Green canary deployments, and administered distributed Cassandra databases.
B.S. in Information Technology · Newark, NJ
Verification ID: F2B4829BC24E13GF
The toolkit behind secure, observable, automated platforms.
Architectures I've designed and shipped end to end.
Architected and deployed an interactive portfolio and AI assistant secured against OWASP LLM vulnerabilities on Kubernetes. Implemented dual-layer guardrails featuring cryptographic canary token monitoring for prompt exfiltration prevention and real-time egress data loss prevention (DLP).
Engineered an isolated execution environment for AI agents and Model Context Protocol (MCP) tools using gVisor container sandboxing with Envoy-based credential redaction.
Ask about my experience, or play red team and try to break the guardrails.
● Online · Gemini on GKE
Suggested questions
⚠ Red team sandbox: test the defenses
Metrics streamed from the running Kubernetes pod, updated every 15 seconds.
No security events triggered yet. Try an injection attack in the assistant above to trigger one!
Need help securing AI workloads, hardening Kubernetes or building a DevSecOps pipeline? Tell me about your project and I'll get back to you within two business days.
I'm open to AI DevSecOps, platform engineering and cloud security roles and consulting engagements.
Warren's AI DevSecOps Portfolio is currently in Private Preview. Please enter your access code to proceed.