10×
Faster releases
GitOps with ArgoCD and health gated rollouts made deployments 10× faster and saved over 100 hours of manual work.
Yaiser Avila RodríguezSenior Site Reliability EngineerBased in Switzerland · Remote worldwide
300+ Kubernetes clusters in production across four clouds, without the noise.
Nearly 5 years running production Kubernetes. At Hydrolix I work on GitOps, progressive delivery and automated remediation for 20,000+ pods. Based in Switzerland, open to remote roles with teams anywhere.
Fig. 1 Clusters under managementAWS · Azure · GCP · Linode
Fig. 2 The fleet, live
One methodology, four clouds, one way to ship.
Every change starts in Git, passes health checks and reaches each cloud the same way. Red dots are deployments, and a node that flashes has just been upgraded. Click a cloud to isolate it.
I remove manual work, so your fleet ships faster, costs less to run and stays calm when it matters.
What I deliverMeasured in production
Results, with the numbers.
−27%
Lower cloud bills
Cut AWS platform cost by 27% and EKS extended support cost by 82.5% through autoscaling, right sizing and a planned upgrade from 1.24 to 1.29.
20+ h
A quieter on call
Automated OOM remediation removed more than 20 hours of toil and 20 to 30 incidents a month.
EMEA
Calm incident command
I led EMEA on call and incident command through peak events such as Black Friday and the Super Bowl.
How we would work togetherThree steps
From first call to a 90 day plan.
- Step 1
Intro call
Thirty minutes on your stack, your pain points and the role. No slides and no pressure.
- Step 2
Technical deep dive
I walk your engineers through real case studies, with the numbers, the trade offs and what I would change.
Example: one commit to the edge, with the numbers - Step 3
A plan for the first 90 days
I bring a written plan for reliability, cost and delivery, based on what I learned from you.
Selected work7 case studies
Problem, what I did, what changed.
Case 1 · Hydrolix · 2024 to present
10×
faster deployments
From kubectl edit to GitOps with ArgoCD
- Problem
- Every cluster was changed by hand with kubectl edit. Nothing was audited, a fix in one region could be silently reverted in another, and infrastructure code was tangled with the application manifests.
- What I did
- I proposed splitting them and moving to GitOps with ArgoCD. The team was at capacity, so we agreed I would work on it for about 30% of my time, outside on call. I built a proof of concept, showed it working and got it approved.
- Outcome
- Onboarding a customer dropped from about two hours to 15 to 20 minutes, every change has an audit trail, and it became the base for later automation such as self healing runbooks.
- ArgoCD
- Kubernetes
- Kustomize
- Pulumi
- GitHub Actions
Case 2 · Hydrolix · 2024 to present
Fleet scale
health gated rollouts
Progressive delivery platform
- Problem
- Upgrades across the whole fleet took weeks of manual coordination, with no safe way to batch, retry or roll back.
- What I did
- I co built a platform on Argo Workflows with regional targeting, automated health checks before and after each batch, a live view of the workflow and scheduled rollouts.
- Outcome
- Rollouts are gated on health and can be rolled back with one click. It became the standard way to upgrade the fleet.
- Go
- React
- Argo Workflows
- ArgoCD
- Prometheus
Case 3 · Hydrolix · 2024 to present
20 to 30
incidents a month eliminated
Automated OOM remediation
- Problem
- The same out of memory alerts paged the on call engineer every month, and the fix was always the same manual change.
- What I did
- I designed a system that reads the signals and proposes the right change through GitOps.
- Outcome
- More than 20 hours of toil and 20 to 30 incidents a month were removed from the rotation.
- Kubernetes
- Prometheus
- ArgoCD
- Go
- Pulumi
Case 4 · Hydrolix · 2024 to present
30+
environments newly monitored
A slow Grafana that exposed a monitoring gap
- Problem
- A customer's Grafana dashboards became very slow. Their queries looked normal and every dashboard was affected. Running df -h in the database pod showed the volume 95% full. It also showed a bigger gap: these customer facing Grafana instances were not part of our monitoring.
- What I did
- I resized the volume with the Postgres operator, then cordoned nodes one at a time and restarted the pods so it came back at the right size. After that I extended our ArgoCD setup with the app of apps pattern to cover the Grafana instances, added alerts for disk use and instance health, and wrote the RCA and a guide for adding Prometheus to their infrastructure code.
- Outcome
- Fixed within one on call shift. Prometheus now covers 30+ environments that had no monitoring, so this kind of issue alerts us before a customer notices.
- Grafana
- Prometheus
- PostgreSQL
- ArgoCD
- Kubernetes
Case 5 · Hydrolix · 2024 to present
Automated
bucket capacity extension
Black Friday capacity, automated afterwards
- Problem
- During a Black Friday, customer buckets and the database catalog reached their read and write limits. Engineers in every region were on call and extended capacity by hand all day to keep the catalog from bloating.
- What I did
- Afterwards I proposed a pipeline and infrastructure code changes to automate that extension, built a proof of concept and presented it.
- Outcome
- The team approved the migration, so a peak day no longer depends on manual work from engineers in every region.
- Infrastructure as code
- CI pipelines
- Capacity planning
- On call
Case 6 · Triggle Spain · 2023 to 2024
27%
lower AWS platform cost
AWS cost optimisation and the EKS upgrade
- Problem
- Resources were over provisioned, and clusters sat on an old version in paid extended support, billed at six times the standard rate.
- What I did
- I led a team of 3 through autoscaling policies, cleanup jobs and right sizing, then upgraded 4 clusters from 1.24 to 1.29 with Terraform and Velero backups.
- Outcome
- Platform cost fell by 27% and extended support cost by 82.5%, and the platform scales better.
- AWS
- EKS
- Terraform
- Velero
- CAST AI
Case 7 · Triggle Spain · 2023 to 2024
1 weekend
to restore production CI
Recovering production Jenkins after a failed upgrade
- Problem
- While testing the EKS upgrade on clusters labelled dev, I found out that the production Jenkins pipeline ran there. The upgrade broke Jenkins and deleted its volume, and production deployments stopped.
- What I did
- I spent the weekend restoring the volume from the Velero backup and recovering Jenkins. I also moved it onto the production management cluster, where it belonged.
- Outcome
- Production CI was back after the weekend. Since then I check what really runs on a cluster, whatever its label says, and take a fresh backup before any destructive change.
- Jenkins
- EKS
- Velero
- AWS
- Kubernetes
Earlier work4 more · openclose
- CI/CD automation and 5× faster buildsTriggle Spain · 2023
- Zero cost on call platform with Grafana OnCallTriggle Spain · 2024
- Docker layer cache in S3, builds from 20 to 5 minutesatSistemas · 2023
- Branch Inspector, 80% shorter CI buildsAccenture · 2022
Platform labSelf built, measured
A Kubernetes platform I built, broke on purpose and rebuilt.
Outside my day job I built a production style platform: GitOps, canary releases on a service mesh, monitoring by default, secrets, backup and restore. I break it on purpose and write down what happened, with the numbers. Every figure here comes from that lab, not from an employer.
Stack
- Kubernetes
- Argo CD
- Argo Rollouts
- Linkerd
- Gateway API
- Prometheus
- Grafana
- Vault
- External Secrets
- Crossplane
- Velero
- Kargo
- OpenTofu
- Cloudflare Workers
- GitHub Actions
The canary that passed a bad release
My own release gate read a perfect success ratio on a server that returned errors, because the mesh proxy counted its own health probes. I found the cause, scoped the metric, added two guards so it cannot fail silently again, and ran it again.
Users exposed: about 5 % for 71 seconds, instead of everyone.
Scope. One meshed namespace. mTLS is observed, not yet enforced.
A platform that rebuilds itself from Git
Only Argo CD is installed by hand. Monitoring, secrets, certificates and Vault come from Git in sync waves and converge by themselves in about 20 minutes, then the full verification passes. A manual change in the cluster is reverted in 2 seconds.
A broken cluster is a rebuild, not an outage of the team.
Scope. One known flake: a certificate race needed one manual retry.
Backups that were actually restored
A backup restored into a freshly created second cluster, and again after a full teardown of the lab, with the data intact. A public status page on Cloudflare probes every five minutes and says unknown when its data is stale.
A lost cluster is not a lost service: the backup outlived the cluster.
Scope. 75 bytes of data, so 13 s is a floor, not an RTO.
How I built itP0 → P8b
Built in phases with exit criteria. Each runbook was written from commands that had just run.
- P0Done
A plan for a first SRE hire
An architecture decision record and a day 1 to 180 plan: listen, agree SLOs, then change on call.
- P2Done
A lab that rebuilds with one command
kind, Prometheus, Grafana, Alertmanager and a sample service, created with one command and checked with another.
- P3Done
Releases gated by metrics
Argo CD, Argo Rollouts canaries, Workflows and Events, Kargo promotion across dev, staging and prod, Vault with External Secrets.
- P4Done
Self service infrastructure and backups
Crossplane claims that create real buckets, and Velero backups to an external store, restored into a clean cluster and timed.
- P5Done
Every runbook run again from zero
A clean room pass on a fresh lab: every command executed, drift fixed, then marked verified.
- P6Done
A public status page at the edge
A Cloudflare Worker with D1, Access and OpenTofu, deployed from GitHub Actions.
- P7Done
Monitoring nobody has to remember
New services are scraped with no monitor written, gaps raise an alert, and undeclared workloads are rejected at admission.
- P8Done
Exact 5 % canaries on a mesh
Linkerd with Gateway API chosen after comparing three options against the lab memory budget.
- P8bDone
The whole base in Git
Monitoring, secrets, certificates and Vault under Argo CD in sync waves, so a rebuild only creates the cluster and Argo CD.
More I can show
SkillStateResult, in my labNext step
- Zero trust trafficProvenPod without a mesh proxy rejected at admissionNext step: Enforce mTLS with an authorization policy
- Capacity and costProvenMesh cost about 0.2 GiB against a 12 GB budgetNext step: Measure under real load
- Observability by defaultProvenNew service discovered in 21 s, no monitor writtenNext step: Run the coverage check in GitHub CI
- Release promotionProvenRelease that failed in dev refused for stagingNext step: A preview environment per pull request
- Infrastructure as codeProvenBucket claim ready in 4 sNext step: Same claim against a public cloud
- SLI and status pageDesignedProbes every 5 min, unknown when staleNext step: Read the SLI after a day of probes
- Operating modelDesignedA day 1 to 180 plan for a first SRENext step: Apply it in a team
Open the live demoThe release gate, its drills and error budget, live on a Cloudflare Worker, free plan.
Experience2022 to present
Four companies, one direction: production.
07.2024 – present
Remote, Switzerland
Hydrolix · Site Reliability Engineer
Streaming analytics platform at petabyte scale, running on a multi cloud Kubernetes fleet of 300+ clusters and 20,000+ pods.
- Operated a multi-cloud Kubernetes production fleet (300+ clusters, 20,000+ pods) across AWS, Azure, GCP and Linode, and coordinated the migration from VMs and manual deployments to secure, automated onboarding of new customers using Argo Workflows, External Secrets and AWS Secrets Manager
- Introduced the platform's GitOps deployment layer on ArgoCD and co-built a progressive delivery platform (Kustomize, Argo Workflows) that automated version rollouts, delivering 10× faster deployments and saving 100+ hours of manual work
- Built the global fleet monitoring system and new coverage for business-critical, customer-facing Grafana instances, extending Prometheus monitoring to 30+ previously unmonitored environments
- Owned EMEA on-call and incident command through peak traffic events such as Black Friday and the Super Bowl, and built automated OOM remediation that eliminated 20+ hours of toil and 20–30 incidents per month
“Yaiser was responsible for the reliability of our multi-cloud Kubernetes fleet, which spans more than 300 clusters across AWS, Azure, GCP, and Linode.”
Marty Kagan · CEO
Reference letter
“He contributed significantly to our move towards GitOps and ArgoCD.”
Jason Turner · Technology leader
Managed Yaiser directly07.2023 – 07.2024
Spain
Triggle Spain · Senior DevOps Engineer (Team Lead)
A tourism platform as a service, running on AWS and Azure.
- Led a DevOps team of 3 engineers supporting a tourism PaaS on AWS and Azure
- Reduced AWS platform costs by 27% and accelerated deployments by 500% through CI/CD automation, right-sizing and scheduled cleanup jobs
- Upgraded Amazon EKS from 1.24 to 1.29 across 4 production clusters, cutting extended-support costs by 82.5%
- Replaced a paid on-call platform with Grafana OnCall at zero cost
“Yaiser was a trusted mentor to junior engineers and fostered a collaborative work environment.”
Francesco Lenahan · COO
Reference letter
“Yaiser led the upgrade of the EKS clusters, a complex task that required extensive planning and management skills.”
Jose Antonio Puertas Pertiñez · Frontend Technical Lead
Reference letter02.2023 – 07.2023
Spain
atSistemas · Senior DevOps Engineer
An airline client, with services on premises and on AWS.
- Maintained on-prem and AWS production services for an airline client, ensuring system scaling and reliability
- Automated deployments using CloudFormation, Ansible and Jenkins
- Identified and implemented preventive measures and remediated outdated legacy systems
- Mentored 2 junior SREs on Python, Bash, Kubernetes, Docker and Datadog
02.2022 – 02.2023
Spain
Accenture · DevOps / Junior Site Reliability Engineer
A banking client.
- Built a GitLab API “Branch Inspector” tool in Python for a banking client, cutting CI build times by 80% and Jenkins EC2 agent usage by 10%
- Operated a DevOps monitoring architecture spanning infrastructure, security, data, development and operations
- Supported the CI/CD toolchain: Jenkins, GitLab, SonarQube and Nexus
- Monitored AWS and Kubernetes workloads with ELK Stack and Grafana, automating tasks in Python, Bash and Go
References01 / 09
What colleagues wrote about me.
Managers, company leadership and teammates, in their own words. Click a face to open their LinkedIn profile and check with them directly.
He maintained the GitOps deployment layer built on ArgoCD and automated the rollout of Hydrolix version releases across the fleet. As a member of the on-call rotation, Yaiser helped support the platform through high-traffic events, including Black Friday and the Super Bowl.
I had the pleasure of managing Yaiser for around 18 months as part of our SRE team. During that time, he helped maintain the reliability of our multi-cloud Kubernetes estate... Above all, Yaiser was always reliable, willing to go the extra mile when needed, and genuinely a great guy to work with.
He has a strong understanding of site reliability engineering and takes a thoughtful, practical approach to keeping the systems running, improving automation and solving issues. Yaiser is someone you can count on especially when things get challenging.
Yaiser demonstrated impressive technical expertise, effectively maintaining both on premises and AWS cloud services. His ability to automate deployments and ensure system scalability was crucial to the success of our projects. He was also a great support to junior profiles.
Alex Curiman
Tech Lead, Epam Neoris
Managed Yaiser directly at atSistemas · August 2024
Verify on LinkedInYaiser was a trusted mentor to junior engineers and fostered a collaborative work environment... he showed great dedication and clarity when dealing with critical issues we experienced following our new product launches... Yaiser was an asset to Triggle, and I highly recommend him for any future roles.
He successfully implemented and maintains ephemeral environments with CodeBuild and ArgoCD, which has allowed the company to improve the efficiency and scalability of its projects... He supervises and maintains the CI/CD of the company's entire complex infrastructure, ensuring the continuous delivery of high-quality software.
Jose Antonio Puertas Pertiñez
Frontend Technical Lead, Triggle
Reference letter · February 2024
Verify on LinkedInYaiser stays calm under pressure, which makes him someone people naturally lean on when things get difficult. He's genuinely curious and digs into problems until he truly understands them, so his decisions are thoughtful and well reasoned. He communicates clearly and without ego.
Tameem Akhtar
Site Reliability, Hydrolix
Worked with Yaiser on the same team · October 2026
Verify on LinkedInYaiser has strong hands on experience with Kubernetes and distributed systems... One of his key strengths is observability. During incidents, Yaiser stays calm and focused, quickly works through complex problems, identifies root causes, and drives issues to resolution while maintaining clear communication with the team.
Jose Augusto Schiavoni
DevOps Engineer, Platform Engineering, Hydrolix
Worked with Yaiser on the same team · October 2026
Verify on LinkedInWhat makes Yaiser truly valuable is his consistency. Regardless of the environment or the circumstances, he brings the same reliable, team first energy. Any SRE or infrastructure team would benefit greatly from having him on board.
Nelson Jaime Galván
DevOps and SRE, Hydrolix
Worked with Yaiser at Accenture and Hydrolix · May 2026
Verify on LinkedIn
Stack and certificationsUsed in production
Tools I use every week.
- Clouds
- AWS · Azure · GCP · Linode
- Containers and orchestration
- Kubernetes · Docker · Helm · Kustomize
- Infrastructure as code
- Terraform · Pulumi · Ansible · CloudFormation
- CI/CD and GitOps
- ArgoCD · Argo Workflows · GitLab CI · Jenkins · SonarQube · Nexus
- Monitoring and observability
- Prometheus · PromQL · Grafana · OpenTelemetry · Vector · ELK Stack · Datadog · PagerDuty
- Secrets management
- External Secrets Operator · AWS Secrets Manager
- Databases
- PostgreSQL · ClickHouse
- Programming and scripting
- Go · Python · Bash · TypeScript
- Practices
- Incident response · Root cause analysis · Postmortems · On call · FinOps · Linux
- Certified Kubernetes AdministratorCertified
The Linux Foundation, 2026
- Certified
AWS Certified Cloud Practitioner
Amazon Web Services
- Certified
Claude Code 101 and AI Fluency
Anthropic Academy, 2026
- Training
HashiCorp Terraform Associate (004)
HashiCorp
- Training
French B1
Language course, currently A2 to B1
Education: Master in Software Development (Assembler Institute), MSc Plant Breeding, MSc Plant Biology, BSc Biological Sciences.
Side projectsMy own infra
What I build and run for myself.
The same habits as at work: monitoring from outside the machine, least privilege, and infrastructure I can rebuild.
Running daily
Discord to Claude Code bridge
Drive a coding agent from a Discord thread
A TypeScript bot that opens a Discord thread per task and streams Claude Code's work back while it runs on my machine. It runs as a launchd agent under my own user, is limited to one workspace by a Seatbelt sandbox, and replaced an earlier design that ran as root and buffered its output.
- TypeScript
- launchd
- Seatbelt
- Redis
- Cloudflare Tunnel
Running daily
Heartbeat monitoring, from outside the machine
A watcher that can notice when the machine dies
A machine cannot report its own death, so the alarm lives elsewhere. A reporter on my Mac edits one pinned Discord message every 5 minutes and sends a beat to a Cloudflare Worker. A cron in the Worker checks the last beat in KV, alerts through a webhook if it is missing, and posts a recovery message when it returns.
- Cloudflare Workers
- Workers KV
- Cron triggers
- Discord webhooks
- launchd
Built in 2024
LinkedIn post scheduler
Personal brand automation on the edge
A small interface for post drafts, and a Cloudflare Worker that reads them from a Turso SQLite database and publishes through the LinkedIn API at the planned time.
- React
- Node.js
- Turso
- Cloudflare Workers
- Wrangler
Private demo on request
Écoute, a French listening lab on local AI
Learn French by listening, with no data leaving my Mac
Paste a French text and a neural voice (Kokoro) reads it sentence by sentence. A local model (Ministral 3 8B on Ollama) prepares the lesson: translation, vocabulary in context, comprehension questions and a chat tutor. In dictation mode the code computes the diff and the model only explains the differences. It runs on an Apple M3 Pro with 18 GB of memory as a small Node server with no npm dependencies, behind a Cloudflare Tunnel and Cloudflare Access.
- Ollama
- Kokoro TTS
- Node.js
- Cloudflare Tunnel
- Cloudflare Access
FAQ8 questions
Questions recruiters ask.
Where are you based?
I live in Fribourg, Switzerland. Swiss employers need no sponsorship because I hold a valid B permit. Everywhere else I work remotely.
Do you work remotely?
Yes. I have worked fully remote on a global team since 2024 and I am on Central European time. Inside Switzerland I hold a B permit, so no sponsorship is needed.
Are you looking for full time roles only?
Full time first. I will consider contracts when the scope is clear and the work is substantial.
What stack do you know best?
Kubernetes with ArgoCD and GitOps, Pulumi and Terraform, Go for tooling, and Prometheus with Grafana for observability.
How big are the environments you have run?
More than 300 production clusters and 20,000 pods across AWS, Azure, GCP and Linode.
Which certifications do you hold?
Certified Kubernetes Administrator (2026), AWS Certified Cloud Practitioner and the Anthropic Academy courses on Claude Code and AI fluency (2026). I am currently preparing for the Terraform Associate (004) exam and studying for French B1.
Which languages do you speak?
Spanish is my native language, my English is fluent (C1) and used daily at work, and my French is A2 to B1 with courses under way, aiming for B1.
When can you start?
My notice period is one month, so I can start one month after we sign.
Contact
Let us talk about your fleet.
Thirty minutes, no preparation needed. I work remotely from Switzerland on Central European time, with teams anywhere. Available one month after signing.
Book a 30 min intro callDownload CVOr email [email protected]