pasindu-thiwanka

LIVE

This portfolio
is running.

You're connected to the control plane of one engineer. Everything on this page is scheduled, observed, and self-healing — because that's what I build: platforms, pipelines, and event streams that stay up so people don't have to.

View workloads
anomaly-detectorself-heal-opsfile-gatewayevent-autoscaleriot-enricherlog-unifiererp-digital-twinci-platformevent-relaycontrolplanepasindu-thiwanka · ctx:production

$ describe operator

Operator

1/1 ready

Manifest

NAME
Pasindu Thiwanka
ROLE
Platform & Systems Engineer
NAMESPACE
pasindu.fyi
MODE
Always reconciling

I run the layer beneath the product — the clusters, pipelines, and event streams everything else stands on. The goal is a substrate so dependable it gets mistaken for physics.

Reconciliation policy

RULE-01 · own-the-substrate

Products get the headlines; platforms carry the weight. I take the box in the diagram that everyone draws and nobody wants to own.

RULE-02 · automate-on-second-occurrence

Anything that happens twice becomes a workflow. Known faults get remediated by machines before a human is paged.

RULE-03 · build-for-the-next-owner

Modular, documented, boring to operate. A system is done when someone who isn't me can run it without a handover call.

$ get workloads --all-namespaces

Workloads

9/9 running

$ get nodes --show-capacity

Node Capacity

9 pools
  • 4/4Runs production traffic
  • 3/4Ships unsupervised
  • 2/4Productive, docs open
  • 1/4Scheduled for evaluation

pool/languages6 units

  • Go4/4
  • Java4/4
  • Python3/4
  • PL/SQL3/4
  • Bash3/4
  • Rust1/4

pool/cloud-infra9 units

  • Kubernetes4/4
  • Docker4/4
  • Helm4/4
  • Terraform3/4
  • Azure3/4
  • AWS3/4
  • Linux/Unix3/4
  • GCP2/4
  • Ansible1/4

pool/backend-data10 units

  • REST APIs4/4
  • Kafka3/4
  • PostgreSQL3/4
  • Oracle3/4
  • MongoDB3/4
  • Redis3/4
  • NATS2/4
  • Hadoop2/4
  • gRPC2/4
  • RabbitMQ1/4

pool/ci-cd6 units

  • ArgoCD4/4
  • Tekton4/4
  • Bitbucket CI4/4
  • Jenkins2/4
  • GitHub Actions2/4
  • GitLab CI2/4

pool/automation4 units

  • Power Automate4/4
  • Selenium3/4
  • UiPath2/4
  • ProcessMaker2/4

pool/observability5 units

  • Prometheus3/4
  • Grafana3/4
  • Elastic Stack3/4
  • OpenTelemetry2/4
  • Jaeger2/4

pool/practices6 units

  • OOP4/4
  • Data Structures & Algorithms4/4
  • Microservices3/4
  • Event-Driven Architecture3/4
  • Machine Learning2/4
  • Domain-Driven Design1/4

pool/collab-tools4 units

  • JIRA3/4
  • Confluence3/4
  • ServiceNow3/4
  • Figma3/4

pool/frontend3 units

  • React2/4
  • Next.js2/4
  • Tailwind CSS1/4

$ get ingress

Ingress

2 routes

Open a route.

Describe the system you need in one paragraph. The reconciler — that's me — takes it from there: the platform, the pipelines, the parts that have to work at 3 a.m. You keep your eyes on the product.

ALL SYSTEMS NOMINALsession 00:00REV D