Homelab
One box, five services
A self-hosted environment on Proxmox — reverse proxy, Git, monitoring, and backups with a recovery path I've actually tested.
Read the write-upAbout
Platform & reliability
I'm a computer engineering graduate working as a junior DevOps engineer in Kathmandu. I spend my time on the layer between a commit and a customer: build pipelines, container images, Linux servers, and the monitoring that tells us when something is wrong.
I like systems the next person can understand — runbooks that are accurate, alerts that mean something, and backups that have actually been restored. Most of this job is making the exciting parts happen less often.
Currently: maintaining internal deployment pipelines, tightening alerting, and taking a larger share of the on-call rotation.
Selected Work
Delivery
A pipeline that tells the truth
CI/CD for this site — preview deploys, cache-aware builds, smoke tests, and a one-command rollback.
Read the write-upObservability
Uptime, measured
SLOs and dashboards for services nobody else watches. Alerts that page on symptoms, not noise.
Read the write-upBackend
A deliberately boring API
A small Go service with structured logs, graceful shutdown, health checks, and a container image under 20 MB.
Read the write-upNotes
Rebuilding the homelab with Terraform
State files, drift, and why "it works on my machine" is not a disaster recovery plan.
A boring CI/CD pipeline is a good pipeline
Preview environments, smoke tests, and a rollback button nobody has needed yet.
Reading Linux logs without panicking
journalctl, grep, and an order of operations for the first five minutes of an incident.
Get in Touch
I'm open to conversations about platform work, reliability, and teams that care about the unglamorous parts. Email is the fastest way to reach me.