Ebenezer Olumeyan

I’m an Infrastructure and DevOps engineer with 10+ years designing and operating mission-critical systems at scale, with a focus on air-gapped, offline, and edge environments where cloud tooling can’t reach. I architect and operate distributed infrastructure end-to-end—planning deployments, automating operations, and keeping reliability measurable.

At Ahmadu Bello University’s Distance Learning Centre, I architected and operate the air-gapped Computer-Based Testing (CBT) infrastructure using 70+ Ubuntu 24.04 LTS servers across 70+ regional locations. At peak, my setup delivered examinations to up to 60,000 candidates per cycle with sustained 90%+ uptime and zero cloud dependency. I also reduced examination delivery time and operational cost by ~30% through Linux server standardisation, configuration automation, and performance optimisation across distributed sites.

I build resilient deployment and operations pipelines for offline reality. I designed Dockerized container environments (Docker Swarm, Traefik) for zero-downtime, standardised deployments—cutting on-site infrastructure provisioning time and cost by ~80%—and implemented CI/CD with GitHub Actions and GHCR, including a staging environment and a manual approval gate for controlled production releases. For real-time insight across the distributed network, I architected observability with Prometheus, Grafana, and Uptime Kuma.

To strengthen operational autonomy in constrained environments, I designed and deployed InfraAgent, an autonomous offline AI incident-response system, using locally hosted LLMs and Ansible-driven remediation with risk gating for operator approval. I also run local, quantised LLM inference with Ollama (GGUF) on constrained hardware, validating viable edge-AI for offline environments, and I’m currently completing my Master of Information Technology (AI specialisation) while building open-source agentic AI tooling for offline infrastructure.