Kubernetes Cost Reduction — AI Security Platform
Cut $350K/year through Karpenter spot scheduling and dangling EBS volume cleanup. No performance degradation, no production incidents.
Karpenter
EKS
Terraform
Staff DevOps & Platform Engineer
Los Angeles, CA
Currently Staff DevOps @ MixMode (AI security platform) — taking select contracts
$0K
AWS savings this year
$0K+
saved at PubNub
0 yrs
in production infrastructure
Platform Engineering
Build and operate EKS clusters, GitOps pipelines, and internal platform tooling. I've run 11+ clusters across prod and non-prod with multi-account ArgoCD.
Cost Optimization
Cut $350K/year at my current company, $750K+/year at PubNub. Karpenter, right-sizing, EBS cleanup, spot adoption — done in production, not in theory.
Observability
VictoriaMetrics, Grafana, AlertManager, Loki. I've deployed this stack at three companies. I also build internal tooling on top of it.
Internal Tooling & Automation
When teams keep doing something manually, I build the tool that automates it. FastAPI, Python, GitLab integrations, AI-assisted workflows.
Sanitized for confidentiality
Kubernetes Cost Reduction — AI Security Platform
Cut $350K/year through Karpenter spot scheduling and dangling EBS volume cleanup. No performance degradation, no production incidents.
Monitoring Stack Replacement — Real-Time Messaging Platform
Replaced existing monitoring tooling with VictoriaMetrics. $750K+/year saved. Deployed to EKS via ArgoCD with full Slack/PagerDuty alerting.
Multi-Cluster GitOps — 11 EKS Clusters
Built centralized ArgoCD setup managing deployments across 11 clusters and multiple AWS accounts using cross-account IRSA. Cluster spin-up went from days to a few hours.
AlertManager Configuration Platform
Built internal web app (FastAPI + HTMX) so engineers can manage AlertManager routes and receivers through a UI — all changes produce GitLab MRs automatically. GitOps workflow without touching YAML.
AI-Assisted Reliability Orchestrator
in progressArchitecting a failure classification and remediation system using LangGraph + FastAPI. Rule-based triage first, LLM reasoning only for anomalies. Structured output with confidence scores and approval gates.
Built and operated solo, outside the day job
ForgeGrip — Multi-Tenant Gym Platform ↗
Designed, built, and operate a multi-tenant SaaS app for gyms — member check-ins, leaderboards, points, and performance tracking. Live in production at a real gym with 200+ active members, mobile-first (93%+ of traffic on phones). Full ownership: schema design, auth and tenant isolation, deployment, and ongoing support — the same scope of responsibility as a platform team, solo.
Contract / Embedded DevOps
3–6 month contracts. I work as a Staff-level IC — I write code, build systems, run the on-call, and ship things. Not advisory, not oversight. Actual work.
Good fit if you need
Rate available on request. Starting with a short discovery call.
Start the conversation →I've been doing this for 12 years. Started at a VoIP company patching ShellShock across 20,000 servers with a Perl script, now I'm running infrastructure for AI security platforms and building AI-assisted reliability tools.
Most of my career has been at companies where I was the DevOps team, or close to it — which means I know how to operate at Staff level without needing a lot of hand-holding. I care about systems that actually work and tooling that makes engineers' lives less painful.
I'm based in Los Angeles. I work remotely.
Want to chat — about a role, a contract, or anything else? Reach out directly, no form to fill out.