DevOps and platform engineer — Kubernetes, ArgoCD GitOps, self-hosted infrastructure
and container security. I write the services running on it too, which is why the
pipelines actually fit them.
Built a self-hosted K3s cluster from bare server to running platform: ArgoCD GitOps with automated prune and self-heal, two applications across separate namespaces, shared stateful infrastructure, autoscaling, disruption budgets, and hard cloud billing guardrails after an account suspension.
Took a live job portal off hardcoded build secrets and onto GitOps: App-of-Apps restructuring, autoscaling and disruption budgets, a secrets pipeline into the cluster, and configuration that auto-deploys — change a key in the secret store, the operator syncs and the deployment rolls itself. Plus the production incidents that came with it.
Ran a company's entire self-hosted stack as a one-person platform team: mail deliverability with SPF, DKIM and DMARC, single sign-on, notifications, self-hosted CI, TLS certificates, reverse-proxy routing, and secrets as the single source of truth — plus the outages and rotations that come with owning it.
A GPU rental platform where untrusted users run their own containers on your hardware. Linux namespace isolation, sidecar-based SSH with no daemon in the workload, per-job egress firewalls, capability dropping, and inter-rental cleanup — then two adversarial pentests against my own design, with all 85 findings closed.
A multi-tenant video SaaS taken from requirements to a containerised, GitOps-deployed production system — migrated off a single-VPS plan onto Kubernetes with ArgoCD, self-hosted secrets and self-hosted object storage, keeping infrastructure cost at the estimate I put in the architecture document.
GoTypeScriptNode.jsJava / Spring BootPythonReact / Next.jsPostgreSQLMongoDBMySQLRedisgRPC
Trajectory
2026, in order
2025 — present
DevOps / Platform Engineer — Techonsy
Design and operate self-hosted K3s clusters with ArgoCD GitOps, delivering multiple production applications through automated, git-driven deployments. Built the shared platform services those applications reuse, led security hardening across them, and authored an end-to-end DevOps training curriculum while mentoring junior engineers.
Jun — ongoing
Platform operations, self-hosted stack
Sole operator of a company's self-hosted services: mail hardening (STARTTLS relay, SPF/DKIM/DMARC, real TLS certificates, removing a vendor backdoor), SSO, notifications, self-hosted CI, and secrets discipline. Autonomous pull-request review across three infrastructure repositories.
Jun — ongoing
K3s cluster + ArgoCD GitOps for two applications
Cluster bring-up, GitOps delivery with prune and self-heal, shared PostgreSQL and Redis in an infrastructure namespace, secret synchronisation via machine identity, autoscaling and disruption budgets, then a storage migration with billing guardrails.
May — ongoing
GitOps build-out and config auto-deploy
Moved secrets out of build specifications into a managed store, restructured ArgoCD into App-of-Apps, and built a pipeline where a configuration change syncs into the cluster and triggers its own rolling restart. Cut a frontend over from CDN to cluster with automated certificates.
May — ongoing
LearnHub — VPS to Kubernetes migration
Re-planned a single-server deployment onto Kubernetes with ArgoCD, self-hosted secrets and object storage; authored the manifests, Dockerfiles and CI workflows, and ran two security audits.
Designed the isolation model for untrusted container workloads — namespace sharing, sidecar SSH, per-job egress firewalls, capability bounding — then attacked it myself across two pentests and closed all 85 findings.
Built and deployed a Spring Boot web application for the real-estate and construction domain — REST APIs, MySQL, authentication, property listings and booking — and contributed to the server setup and deployment behind it.
2026 · 2024
M.Sc. & B.Sc. Computer Science — Sir Parashurambhau College
M.Sc. with 9.14 CGPA (2026), B.Sc. with 8.81 CGPA (2024).
About
How I operate
I'm a DevOps and platform engineer. Most of my work is the part that happens after the merge button:
cluster bring-up, GitOps delivery, secrets, certificates, networking, and being the person who gets
paged when it breaks. I came to it from application development, which is why I read a service's code
before I write its deployment.
The work I'm proudest of is unglamorous. A kubelet crash-looping because fstab re-enabled
swap on a 2 GB server after every reboot. A 413 on uploads that turned out to be the
frontend's nginx catch-all, not the backend. A "networking" outage that was really a Postgres
password failure — found only after ruling out firewall drops and SSH forwarding in order.
Root cause over workaround, every time.
I default to self-hosted and cost-conscious infrastructure over managed cloud, keep secrets in a
managed store rather than in build config, and treat a hard billing cap as better than a surprise
invoice. I write the runbooks and threat models as I go, because infrastructure nobody else can
operate is a liability, not an achievement. That's also why I authored an end-to-end DevOps
training curriculum and mentor the junior engineers on my team — the platform has to outlive
whoever built it.