Site Reliability Engineer Blog/Portfolio
SRE with 5+ years across the cloud-native and consulting industries. Specialised in GitOps, multi-cloud Kubernetes at scale, cost optimisation, and automation.
About Me
I’m Yaiser Avila Rodríguez, a Site Reliability Engineer with 5+ years building and operating large-scale, multi-cloud Kubernetes infrastructure across AWS, GCP, and other providers. Currently working on multi-region production fleets in a cloud-native streaming-analytics environment.
My focus is on turning manual ops into automated, observable systems — from GitOps adoption and progressive delivery platforms to fleet-wide cost-optimisation initiatives that have delivered measurable savings and reclaimed significant engineering time.
Outside work I run InfraBio, a personal brand where I share SRE content and tooling.
Work Experience
Site Reliability Engineer
Hydrolix
Remote
- ▸Drove a Vertical Pod Autoscaler rollout across the production Kubernetes fleet using a phased, region-gated approach
- ▸Led company-wide GitOps adoption with ArgoCD as the unified continuous delivery model
- ▸Co-designed and built a fleet-wide progressive delivery platform on ArgoCD + Kustomize + Argo Workflows — health-gated, automated rollouts with one-click rollback
- ▸Built and operated Prometheus + Grafana observability for production environments
- ▸Designed an automated OOM-remediation system that surfaces GitOps changes automatically based on observed signals
- ▸Owned EMEA on-call and incident command; significantly reduced triage time using AI-assisted tooling; supported capacity planning for major scheduled events
- ▸Sustained high-volume contribution cadence across code, internal tooling, and technical documentation over 2 years
DevOps / Cloud Engineer
Triggle Spain SLU
Spain
- ▸Upgraded EKS clusters 1.24 → 1.29, reducing extended support costs by 82.5%
- ▸Led 27% AWS platform cost reduction via autoscaling policies, cleanup crons, and resource right-sizing
- ▸Managed a team of 3 engineers automating key infrastructure components
- ▸Implemented Grafana OnCall + Prometheus alerting, replacing paid UptimeRobot subscription at zero cost
DevOps Engineer
Knowmad mood
Spain
- ▸Reduced Jenkins build times from 20 min to 5 min (75%) by integrating Docker layer cache stored in S3
- ▸Implemented S3 cleanup lifecycle policies to control cache storage costs
DevOps Engineer
Accenture
Spain
- ▸Built Branch Inspector tool using GitLab API + Python — cut CI build times by 80%
- ▸Reduced EC2 Jenkins agent usage by 10% through stale branch detection and automated cleanup alerts
- ▸Integrated Kubernetes ConfigMaps, Logstash, Prometheus, and Grafana dashboards for branch monitoring
ls ~/sandbox/
live SRE demos · running in your browser · no backend
Est. Resource Waste
mid fleet · CAST AI: 99.94% over-provisioned · Datadog 2024: 83% idle · Kubecost: 35–50% baseline
35%
auto-calculated
VPA/HPA Implementation
VPA 25.9% + HPA 7.5% of bill · 150 clusters
GitOps Automation
3 SREs × 30 hrs/release × 12 · 150 clusters
Alerts / Runbooks / Auto-Remediation
5% auto-resolved/mo · 35 min · 150 clusters
Projected Annual Savings
$287Kvs. 1 SRE-month investment ($16K)
payback in ~2.9 weeks
CAST AI · Datadog · Kubecost benchmarks
Year-1 financial model
ROI
1.8×
self-funds in
~29wk
What you actually buy
“Don’t hire an SRE to react to incidents. Hire one to build the systems that prevent them — and fund their own salary while doing it.”
Traffic
Request Rate
Errors
Error Rate
Latency
P99 Latency
Saturation
CPU Saturation
synthetic data · real Prometheus patterns · 4 SRE golden signals (Latency · Traffic · Errors · Saturation)
Flagship Projects
Triggle Spain SLU (2024)
EKS Upgrade 1.24 → 1.29 — 82.5% Extended Support Cost Reduction
Problem
EKS clusters were running version 1.24, which had entered extended support — priced at 6× the standard rate. With no upgrade plan in place, costs were compounding monthly.
Solution
Oversaw the full upgrade path from EKS 1.24 to 1.29 using Terraform and Velero for workload backup. Managed a team of 3 engineers automating key infrastructure components and autoscaling group configurations throughout the process.
Outcome
82.5% reduction in extended support costs. Enhanced system scalability and reliability post-upgrade.
Technologies Used
Terraform, Terraformer, Velero, Kubernetes (EKS), AWS Auto Scaling
Triggle Spain SLU (2023)
CI/CD Automation and Deployment Acceleration
Problem
The CI/CD process was entirely manual, prone to human errors, and slow — delaying the deployment of new features to production.
Solution
Automated the full CI/CD pipeline with concurrent builds and Docker layer caching. Overhauled deployment methodologies and introduced automation scripts that streamlined the end-to-end deployment process.
Outcome
Deployment and build times reduced by 5×. Enhanced consistency and reliability across all deployments.
Technologies Used
AWS CodeBuild, Bitbucket, Lambda, API Gateway, Python, Bash
Triggle Spain SLU (2024)
AWS Platform Cost Optimisation — 27% Reduction
Problem
Over-provisioned resources, suboptimal allocation, and the lack of effective autoscaling and cleanup policies were generating significant unnecessary AWS spend.
Solution
Led a cost reduction initiative across AWS accounts: enhanced autoscaling capabilities, implemented cleanup lifecycle policies, automated resource cleanup via crons, and revised resource allocation to match actual usage patterns.
Outcome
27% reduction in overall platform costs via autoscaling and cleanup policies. Improved budget efficiency and resource utilisation.
Technologies Used
AWS, ArgoCD, CodeBuild, Terraform, CAST AI
Tools & Technologies
Continuous Delivery
Cloud & Infrastructure as Code
Observability & Reliability
Automation & AI
Let’s work together
Open to new roles, collaborations, or a conversation about SRE, Kubernetes, and infrastructure automation.