Selected Work

Personal systems and product prototypes covering safer deployments, monitoring, infrastructure automation, and day-to-day operational problems. Each project is labeled by its current status.

View GitHub Repositories
Your Signal project screenshot

Your Signal

URL

Personal uptime-monitoring service with a blue/green deployment pipeline. An idle systemd instance is updated and health-checked before Caddy switches traffic; a database-backed leader lease coordinates monitoring jobs to avoid duplicate checks during deployment. Currently used internally.

  • Internal use
  • Release engineering
  • Go
ASMT project screenshot

ASMT Dashboard

URL

Open-source Go dashboard for monitoring Linux servers, focused on service health, log visibility, and fast operational status review from a lightweight agent.

  • Open source
  • AGPL-3.0
  • Observability
WaterFire project screenshot

WaterFire

Personal-use Schwab autotrader running on GCP, built in Python with SQLite and a state-driven scheduler. AI is limited to sentiment refinement and tie-break decisions.

  • Personal use
  • GCP
  • Scheduled workloads
Meemyo project screenshot

Meemyo

URL

Personal product prototype: a speed-first Go CMS with RBAC, server-rendered themes, reusable content modules, and an asynchronous image pipeline.

  • Prototype
  • Go
  • Product design

Home Observability Lab

Active personal system

  • Run a home lab for software development with VMs hosted on TrueNAS, an OPNsense router/firewall, and an Intel NUC. Deploy the central monitoring stack through Docker Compose.
  • Collect logs and metrics with Grafana Alloy, rsyslog, and node/process exporters feeding Loki and Prometheus. Grafana dashboards cover host health and network activity.
  • Review OPNsense firewall logs for suspicious traffic and report suspicious source IPs through the AbuseIPDB API.
  • Use authenticated HTTPS over LAN/WireGuard and source-restricted log ingestion. The dashboards have helped identify post-reboot monitoring failures, temporary DNS failures during upgrades, and misleading startup alerts.