Hi, I am

Alejandro.

Senior Platform & Site Reliability Engineer · Multi-Cloud (AWS + GCP) · Kubernetes · LLM Infrastructure

I build the infrastructure that engineering teams ship on. The piece I’m proudest of is the AWS platform behind Colombia’s first fully digital bank, which we ran at a 99.999% availability SLO. Over 15+ years in finance, fintech, government, and media I’ve also run self-hosted LLM inference that kept regulated data in-house, moved production between AWS and Google Cloud, and led DevOps, SRE, and Cloud teams of up to 25. I go deep on Kubernetes and Terraform, and I keep the SRE habits that make them pay off: SLOs, disaster recovery, on-call. Based in Medellín, Colombia, working remotely across US and EU hours.

View Selected Work
Alejandro Llanes — Platform & Site Reliability Engineer
15+
Years in production infrastructure
99.999%
Availability SLO, digital bank
~25
Services migrated AWS→GCP, zero-downtime
5-min
RPO on validated bank DR drills

Selected Work

A few things I’ve built, anonymized. Happy to walk through the architecture on request.

Colombia's first digital bank, run at a 99.999% SLO
AWS DynamoDB Global Tables Aurora MySQL Terraform AWS DRS FinOps
Colombia's first digital bank, run at a 99.999% SLO
The AWS platform behind the country's first fully digital bank, run to a 99.999% availability target with real SLIs and an error-budget policy. Multi-region HA/DR (DynamoDB Global Tables, Aurora MySQL, cross-region S3, AWS DRS), proven in failover drills at a 5-minute RPO and 4-hour RTO. Peak compute cost came down about 30%, and releases went out daily at a 15%-or-lower change-failure rate.
Read the write-up
Self-hosted LLM inference that kept regulated data in-house
vLLM Ollama FastAPI EKS LangChain Amazon Bedrock
Self-hosted LLM inference that kept regulated data in-house
A self-hosted LLM inference platform for a regulated bank, so its data never had to leave the building. Open-weight models (vLLM/Ollama) on GPU-backed EC2 plus Amazon Bedrock and hosted APIs, all behind FastAPI OpenAI-compatible endpoints and OpenWebUI on EKS. It ran under the same IaC, CI/CD, secrets, and security controls as everything else, with LangChain handling retrieval over application logs and operational data.
Read the write-up
Zero-downtime AWS→GCP re-platform of ~25 services
GCP (GKE) AWS Terraform GitLab CI DMS Memorystore/Redis
Zero-downtime AWS→GCP re-platform of ~25 services
Moved a US retail-and-delivery startup's full stack, roughly 25 services, from AWS to Google Cloud (GKE, managed databases, Memorystore/Redis, Cloud Functions, API Gateway, load balancers). The cutover happened incrementally with no downtime, seeded by backup-and-restore and kept current with DMS replication, on a Python and Go stack with GitLab CI and Terraform.
Read the write-up
Cross-border payments DR at a 1-hour RTO and 5-minute RPO
AWS Aurora PostgreSQL ECS Transit Gateway CloudFormation GitHub Actions
Cross-border payments DR at a 1-hour RTO and 5-minute RPO
A multi-region disaster-recovery platform on AWS for a US money transmitter moving funds to Latin America. Built on Aurora PostgreSQL, DynamoDB, ECS with Service Connect, API Gateway, and multi-VPC networking through Transit Gateway, targeting a 1-hour RTO and 5-minute RPO. Delivered as CloudFormation with GitHub Actions CI/CD.

About

I’m a Platform and Site Reliability Engineer, and I’ve spent 15+ years building the platforms other engineers build on. That work has been across AWS and Google Cloud, in finance, fintech, government, and media. The project that taught me the most was the AWS platform behind Colombia’s first fully digital bank. We ran it to a 99.999% availability SLO, with multi-region disaster recovery, self-service environments, and CI/CD that teams could actually trust. I also stood up a self-hosted LLM inference platform (vLLM/Ollama) so the bank could use models without regulated data ever leaving the building. More recently I moved a US startup’s production onto Google Cloud (GKE, Cloud Functions) in a gradual, zero-downtime cutover of about 25 services.

My bias is toward paved roads: reusable Terraform modules, GitOps delivery, and SLOs with error budgets, so product teams can move fast without babysitting infrastructure. I’m certified on Kubernetes and IaC (CKA, CKAD, Terraform Associate), but the certificates matter less than the habits behind them: on-call, incident response, blameless postmortems, real DR drills. I’ve led DevOps, SRE, and Cloud teams of up to 25. I’m based in Medellín, Colombia, and open to remote Platform Engineering, SRE, DevOps, or AI-infrastructure roles across US and EU hours, at Staff/Principal or senior IC scope.

What I work with day to day:
  • Cloud & IaC: AWS (EC2, EKS/ECS, Lambda, RDS/Aurora, DynamoDB, S3), Google Cloud (GKE, Cloud Functions, API Gateway), Azure, Terraform, Pulumi, CloudFormation
  • Platform & CI/CD: internal developer platforms, self-service provisioning, golden paths, reusable Terraform modules, environment promotion, Jenkins, GitHub Actions, GitLab CI, Argo CD (GitOps), canary and blue-green deployments
  • Kubernetes & containers: Kubernetes, Docker, Helm, service mesh (Istio, Linkerd), Kubernetes RBAC
  • Observability & reliability: Prometheus, Grafana, OpenTelemetry, Datadog, New Relic, ELK, SLOs and error budgets, incident management and RCA, disaster recovery and HA, FinOps
  • Security & DevSecOps: SAST/DAST, OWASP Top 10, SIEM, WAF, vulnerability management (Nessus, Qualys, Dome9), IAM, zero-trust segmentation (Calico), mTLS, secrets management (Vault, AWS Secrets Manager, SOPS), SOC 2, PCI DSS, ISO 27001
  • AI Infrastructure & LLM serving: self-hosted LLM inference platform (vLLM, Ollama on GPU-backed EC2), Amazon Bedrock, OpenAI-compatible model gateway and serving (FastAPI), hosted model APIs (OpenAI, Hugging Face, DeepSeek), OpenWebUI, LangChain RAG over logs and operational data, LLMOps/MLOps, AI platform engineering
  • Data & databases: PostgreSQL, MySQL, Amazon Aurora, DynamoDB, Redshift, MongoDB, Apache Airflow, Spark, Amazon EMR, Kafka, Glue, DMS
  • Languages & OS: Python, Go, Rust, Bash, Linux, Unix, FreeBSD
  • Spoken: Spanish (native), English (professional working proficiency)

Experience

Senior Cloud Architect (Contract) — Transnetwork (Movantis), Remote - Transnetwork (Movantis)
Dec 2025 - Jul 2026
  • Architected multi-region disaster recovery on AWS to a 1-hour RTO and 5-minute RPO: Aurora PostgreSQL, DynamoDB, and ECS with Service Connect behind API Gateway.
  • Wired the multi-VPC networking through Transit Gateway and delivered the whole platform as CloudFormation, with GitHub Actions running CI/CD.
GCP Senior Platform Engineer & SRE Consultant — US clients, Remote - Independent · Contract
Jun 2025 - Jan 2026
  • For a US online retail and delivery startup, moved the full stack, about 25 services, from AWS to Google Cloud (GKE, managed databases, Memorystore/Redis, Cloud Functions, API Gateway). The cutover was incremental and took no downtime, seeded by backup-and-restore and kept in sync with DMS replication.
  • Built the same startup a self-service platform on Amazon EKS with isolated DEV/STG/PROD environments and a clear promotion path, replacing Docker-on-EC2; ran delivery with GitLab CI and Terraform across a Python and Go stack (PostgreSQL/PostGIS, Redis, Firebase, Stripe billing).
  • For a US cybersecurity company (AI-driven asset scanning and vulnerability discovery), automated ephemeral per-tenant scanner infrastructure on GCP: about 300 VMs provisioned on demand, 3 per tenant across roughly 100 customers, through Python Cloud Functions behind API Gateway. No standing infrastructure, no manual setup.
DevOps & Infrastructure Lead (Platform Engineering & SRE) — Lulo Bank (GNB Sudameris), Bogotá - Lulo Bank
Jan 2020 - Jun 2025
  • Grew and led the bank’s DevOps, SRE, and Cloud organization to 25 engineers at peak. That meant hiring, mentoring, owning the platform roadmap and budget, and running on-call, incident response, and blameless postmortems.
  • Built the compliance-technology controls, processes, and regulator-facing evidence that supported approval of the bank’s operational banking license. Drove SonarQube (SAST) and Dome9 remediation with the cybersecurity team.
  • Cut peak compute cost by about 30% through FinOps rate optimization (Savings Plans, Spot) while operating the platform against a 99.999% availability SLO target, with instrumented SLIs and an enforced error-budget policy.
  • Validated a 5-minute RPO and 4-hour RTO in live multi-region failover drills: DynamoDB Global Tables, Aurora MySQL, S3 cross-region replication, and AWS Elastic Disaster Recovery.
  • Shipped daily production releases at a change-failure rate of 15% or lower, with sub-hour recovery, through GitOps on EKS (Argo CD, Helm, Linkerd) plus Jenkins, Terraform, and GitHub Actions.
  • Designed and ran the bank’s self-hosted LLM inference platform, keeping regulated banking data in-house: open-weight models (vLLM, Ollama) on GPU-backed EC2 plus Amazon Bedrock, behind an OpenAI-compatible model gateway (FastAPI) and OpenWebUI on EKS. Ran LangChain RAG over application logs and operational data, and wrote internal Kubernetes tooling in Python and Rust.
Head of DevOps — Alkanza US, Bogotá - Alkanza US
Oct 2016 - Nov 2019
  • Built the company’s DevOps, DevSecOps, and infrastructure practice from zero, leading a team of 7 across infrastructure, DevOps, and security.
  • Designed and ran a multi-region AWS platform, migrating standalone containers to auto-scaling Kubernetes with zero-trust Calico networking and mTLS through the Istio service mesh.
  • Took delivery to on-demand deploys with sub-hour lead time and recovery, using Jenkins CI/CD, Terraform and CloudFormation, and Ansible/Puppet/Chef, for microservices on AWS (EKS/ECS/Fargate, Lambda, API Gateway; Aurora, DynamoDB).
  • Built the data pipelines behind analytics and ML workloads: Apache Airflow, Celery, Spark on EMR, Kafka streaming, and Glue/DMS ETL into Redshift Spectrum.
  • Put security in the pipeline (SAST/DAST, OWASP Top 10) with continuous vulnerability management (Nessus, Qualys, Dome9) and centralized secrets (HashiCorp Vault, AWS Secrets Manager, SOPS), and ran ELK logging and SIEM to support SOC 2, PCI DSS, and ISO 27001.
Infrastructure, Systems & Database Administration - Earlier Experience
2008 - 2016
  • Eight earlier roles in infrastructure, systems, and database administration: virtualization (VMware, Hyper-V, KVM), Java middleware (JBoss, Tomcat, GlassFish), Linux and FreeBSD, and PostgreSQL/MySQL/SQL Server high availability and replication.
  • Employers in finance, media, and government, including Zürich de Occidente, PayU Latam, OpenBox LTDA, Diario La República, and Masaro Solutions.

Certifications

Certified Kubernetes Administrator (CKA) badge
Certified Kubernetes Administrator (CKA)
The Linux Foundation · earned Sep 2025. Verify on Credly.
HashiCorp Certified: Terraform Associate (003) badge
HashiCorp Certified: Terraform Associate (003)
HashiCorp · earned Sep 2025. Verify on Credly.
Certified Kubernetes Application Developer (CKAD) badge
Certified Kubernetes Application Developer (CKAD)
The Linux Foundation · earned Oct 2025. Verify on Credly.

Get in touch

If you’re building regulated or high-reliability platforms and want someone who’s actually run one at 99.999%, I’d like to hear about it. LinkedIn is the fastest way to reach me. Send a note and we’ll go from there.
Connect on LinkedIn