Austin, TX · U.S. Permanent Resident, no sponsorship required
12+ years running Kubernetes, Terraform, and GitOps platforms for telecom, government, and enterprise workloads.
Advanced SRE for Staff Engineers
Built on questions I carried through 16 years of running mission-critical telecom platforms. Most serious reliability problems do not come from missing tools. They come from broken feedback loops.
They grow an outage and hold it in place. Retry storms that backoff and jitter cannot fix. Outages that outlive their trigger, where you roll the bad deploy back, traffic returns to normal, and the system is still on fire.
Measurement that never turns into a decision. Error budgets that run out while the launch ships anyway. Postmortems that find the root cause, close every action item, and still let a similar incident happen again.
A technical loop keeps an outage from ending. An organizational loop makes it happen again.
Fourteen chapters covering blast radius through cells and shuffle sharding, telemetry cost and gray failure, capacity by queueing theory, safe config delivery, incident command, and a side-by-side reading of 16 public postmortems.
Platforms and automation I designed, built, and operated in production.
Self-service deployment workflows that let product teams provision their own namespaces, pipelines, and observability without filing a ticket with the platform team. Templated golden paths replaced one-off environment requests, and GitOps made every change reviewable and reversible.
Reusable modules for AWS network, compute, and Kubernetes provisioning, published so every team builds environments the same way. New-environment setup dropped from days to hours.
GitLab CI with self-hosted runners on Kubernetes for elastic, isolated build capacity, and Argo CD for continuous delivery. Jib built images straight from Gradle and Maven, which removed privileged Docker builds from the cluster and cut image build times.
GPU-backed containerized LLM and Jupyter workloads on Kubernetes, exposed through FastAPI and Streamlit endpoints to support AI-assisted engineering workflows and internal developer productivity tooling.
Enterprise monitoring across Datadog, Prometheus, Grafana, and OpenSearch to shorten incident detection and response, paired with FinOps-driven rightsizing that improved cloud resource utilization.
Recent public builds. Each one exists to remove a specific class of failure I have been paged for.
Terraform reference lab with guardrails at four layers: S3 remote state with native locking and CI concurrency groups, apply only from saved plan files with daily drift detection, per-environment state split with for_each keyed modules, and a double gate of variable validation plus OPA/conftest at plan time.
Twenty ordered labs from a first S3 bucket through modules, environment separation, remote state, and a GitHub Actions pipeline, then on to security scanning, state recovery, policy as code, multi-region, IAM least privilege, native Terraform tests, drift detection, ECS, and EKS.
One project context generates nine artifacts in sync: Dockerfile, Compose, Kubernetes manifests, Helm chart, Terraform, GitHub Actions, GitLab CI, Ansible playbook, and Nginx config. Change a port once and every artifact follows, so the config drift behind so many deployment incidents cannot occur.
Shift-left DevSecOps toolkit fusing two workflows over one codebase: a headless scanner with SARIF and JSON output and CI exit-code gates, and a local intercepting proxy with repeater and fuzzer for hands-on verification.
A second brain that is only Markdown, Git, and Claude. GitHub Pages publishes the vault as a dashboard, graph, and task board, and a button on the site opens a GitHub issue that triggers Claude Code inside Actions to write and commit notes. The browser holds zero secrets; execution happens only on the runner.
Turns any network into an interactive knowledge graph and only trusts what randomness cannot explain. PageRank, HITS, betweenness, and Louvain communities are each validated against a degree-preserving null model with Benjamini-Hochberg correction, then rendered by a dependency-free D3 page.
Live knowledge and automation systems I design, build, and operate end to end, from content pipelines to GitHub Actions delivery and custom domains.
Also running: architectures, checklists, automations, comparison, comp-thinking, insight, meetup, and more on GitHub.
Camino, Seoul
Led cloud infrastructure assessments and governance for government cloud-native environments, evaluating security posture, platform readiness, and operational risk. Built Python tooling for infrastructure diagnostics and platform troubleshooting, and partnered with engineering, security, and compliance teams to set cloud governance standards.
SK Telecom, Seoul
Operated enterprise Azure and hybrid cloud infrastructure behind mission-critical telecom platforms at 99.99% availability. Architected Kubernetes-based platforms, standardized Terraform and manifest-driven provisioning, modernized seven CI/CD pipelines onto GitOps, and built the observability and self-service workflows product teams used every day.
SK Telecom, Seoul
Led platform readiness for large-scale distributed systems, including performance, stress, and endurance testing in production-equivalent environments. Authored the runbooks and validation frameworks that reduced post-deployment incidents.
SK Telecom, Seoul
Ran 24x7 enterprise Linux infrastructure serving millions of subscribers. Led major incident response across more than thirty enterprise outages, and drove high-availability, disaster recovery, and capacity planning work.
Led end-to-end development of an enterprise endpoint security management solution, establishing centralized visibility and control over diverse endpoint assets. Security audits and risk assessments applied zero-trust principles to harden the posture.
Directed static analysis with Fortify to surface structural risks, and established a secure test bed that validated both functional and security parameters of cross-unit interactions.
Led a task force standardizing the IT stack across OS, database, WAS, and network, authoring integrated best-practice guidelines that reduced integration-related failures.
Owned monthly KPI reporting for integrated systems, using the resulting data to optimize maintenance cycles.
Resident overseas assignment directing telecommunications network operation KPIs for Ho Chi Minh City and Hanoi, handling technology transfer and quality compliance with local operating teams.
