I build AI-assisted platform engineering tools: containerized Claude agents, MCP servers, and Kubernetes/SRE automation.
My work sits at the boundary between infrastructure and AI-assisted operations: isolated agent runtimes, developer workflows, cluster-aware SRE tools, and small systems that are easy to inspect and operate.
- Containerized AI agents with explicit filesystem, process, and credential boundaries
- MCP servers that connect AI tools to operating-system and developer workflows
- Kubernetes observability tools for discovery, dashboards, alert tuning, and incident investigation
- CLI and automation tooling for Docker, homelab, and platform engineering workflows
| Project | What it proves |
|---|---|
| MetaClaw | Containerized Claude agents with per-group filesystems, SQLite state, IPC, scheduled tasks, and explicit mount validation |
| KubePulse | Kubernetes health and observability tooling across a Go backend, React frontend, CI, Docker Compose, and cluster checks |
| notifications-mcp-server | A focused MCP utility with macOS notifications, Claude Desktop configuration, tests, and repeatable local verification |
| Project | Area | Notes |
|---|---|---|
| MetaClaw | Containerized agents | Node.js orchestrator for Claude agents running in isolated containers, with messaging channels, SQLite state, IPC, scheduled tasks, and mount validation |
| KubePulse | Kubernetes monitoring | Go and React Kubernetes health monitor with cluster checks, alerting/SLO modules, AI diagnostics, Docker Compose, tests, and CI |
| sre-plugin | Kubernetes observability | Claude Code SRE tooling for cluster discovery, service dashboards, Prometheus/Loki/Grafana context, and alert investigation |
| notifications-mcp-server | MCP | MCP server that exposes native macOS notifications to Claude Desktop and related agent workflows |
| auto-pr | DevTools CLI | Go CLI that uses Claude Code plus gh/glab to generate pull request and merge request text from local git context |
| claude-code-sub-agents | Claude Code workflow | Lightweight installer for an agent-architect that creates project-specific Claude Code agents |
AI platform and agent systems
- Claude Code, Claude Agent SDK, Model Context Protocol
- Agent orchestration, per-session state, tool routing, IPC, and scheduled execution
- Container isolation for agent workloads using Docker and Apple Container
Platform engineering
- Kubernetes, Docker, Compose, Helm, Terraform
- CI/CD with GitHub Actions and related automation
- Linux/macOS service management, local infrastructure, and secure remote access
Observability and reliability
- Prometheus, Grafana, Loki, Alertmanager
- Cluster discovery, SLOs, dashboards, alert review, and incident workflows
- SQLite-backed state, recovery paths, retries, and operational debugging
- Keep the core system small enough to audit
- Prefer explicit trust boundaries over broad permissions
- Use automation to remove repetitive operational work
- Document architecture, security assumptions, and known limitations
Email: charles.adedotun8@gmail.com



