SYSTEM RECONSTRUCTION
Banco Limen
A Phase 1 reconstruction of legacy banking failure patterns: its intentional anti-patterns, decision records, and the reasoning that will guide each later fix.
Explore LimenPlatform engineering from the systems up
Reliable platforms begin where abstractions end.
I’m Cristian Gómez Aranda — a Platform Engineer with eight years operating critical banking systems. I trace Linux and Kubernetes internals to build infrastructure that teams can trust in production.
Selected impact
- ~22.5K
- estimated annual manual hours removed through automation
- 4H 15M
- a critical recovery path reduced with IaC and observability
- 8 YEARS
- production systems experience at Citibanamex
NEXT: foundations for secure, observable AI workloads.
TECHNICAL WRITING
From Scratch
A public series that builds Kubernetes networking from Linux primitives — namespaces, veth pairs, iptables, conntrack, and the packet path in between.
Start readingWHAT’S NEXT
AI workload infrastructure
A next area of study: applying the same operational discipline to secure, observable infrastructure for model serving and AI teams.
Why this directionFrom Splunk to Kubernetes: what 8 years of alert fatigue taught me about self-healing systems
Eight years running production SRE at a major bank taught me one thing about alert fatigue: detection is the hard part. Kubernetes changed that assumption for me today.
Exposing a service without Kubernetes: DNAT load balancing with iptables
How to expose a container to the outside world and distribute traffic across multiple backends using iptables DNAT — the same mechanism kube-proxy uses for NodePort and ClusterIP services.
How pods reach the internet: NAT and masquerading
Why pods can't reach the internet by default and how iptables MASQUERADE fixes it — a deep dive into NAT, conntrack, and the path packets take out of a Kubernetes node.
Building a container network from scratch: namespaces, veth, bridges
How Kubernetes pod networking actually works — built from Linux primitives: network namespaces, veth pairs, and bridges. No abstractions, just the kernel.
Open to remote Platform Engineer and SRE roles with teams running Kubernetes in production — based in Mexico City, strong overlap with US business hours.
Start a conversation