Meshari AlDoweesh
Site Reliability Engineer — Riyadh
Site Reliability Engineer with hands-on fintech experience operating Kubernetes workloads, provisioning cloud infrastructure with Terraform, and building CI/CD, GitOps, observability, and load-testing workflows. End-to-end experience deploying and operating production systems with automated backups, restore testing, monitoring, and zero-trust access.
Experience
- —Built a shared platform development environment from scratch using Terraform, mirroring the company's staging architecture and supporting engineering teams' daily development workflows.
- —Automated ClickHouse deployment using Ansible as part of the company's analytics-platform evaluation; the team projected approximately SAR 1M in annual savings compared with BigQuery, and ClickHouse was subsequently adopted as the production analytics store.
- —Prototyped an internal infrastructure-investigation assistant integrating Prometheus, Grafana, and Alertmanager, designed for Helm-based deployment on Kubernetes.
- —Built k6 load-testing suites for critical fintech user journeys: registration, authentication, OTP verification, transactions, and wallet operations.
- —Deployed and managed microservices on production Kubernetes clusters via GitOps (Argo CD, Flux CD), including Helm chart customization and GitLab CI/CD pipelines.
- —Provisioned OCI infrastructure through the internal Terraform module repository and supported live production deployments.
Projects
One command from an empty AWS account to a production-grade k3s cluster: Terraform-provisioned, GitOps-managed by Argo CD (app-of-apps, prune + self-heal), TLS ingress, Prometheus/Grafana/Loki observability, and tested etcd backups.
AI incident-triage agent for Kubernetes: when an alert fires it investigates through bounded read-only tools and posts a diagnosis with cited evidence to Telegram. Read-only by design — the human stays in command.
Deployed and operate a production web application solo: DigitalOcean, Docker Compose, PostgreSQL, Cloudflare Tunnel + Zero Trust (zero inbound ports), automated encrypted backups with restore testing, and uptime monitoring.
Multi-model AI platform with unified memory: integrates multiple LLM providers with embedding + recency ranking for contextual retrieval and user-controlled memory (review, correct, delete).