Self-serve by default
Ship a values schema and an install button, not a template and a wiki page.
DevOps Engineer
ARUNESH J
I build the multi-cloud platform other engineers deploy on — four clouds, one self-serve path.
status
operational
region
Bengaluru · IN
since
2024 · 2y+ uptime
local
00:00:00
Platform Engineering
Kubernetes
Helm
Terraform
OpenTofu
Go
Multi-Cloud
Infrastructure as Code
AWS
GCP
Azure
OCI
Helm Charts
Internal Developer Platform
Self-Serve Provisioning
GitHub Actions
CircleCI
Docker
GoFr
REST APIs
Microservices
Redis
Grafana
Prometheus
Loki
Mimir
Observability
CI/CD
DevOps
Site Reliability Engineering
Cloud Cost Optimization
Values Schema
CEL
Production Readiness
Release Engineering
Python
Bash
Linux
Git
SQL
Vertex AI
AWS Bedrock
ONNX
GKE
Managed Databases
Networking
Incident Response
Configuration Drift
Open Source
01 — whoami
A platform is a product.
A platform is a product.
I ship it like one.
Two years collapsing cloud setup into a 30-minute, self-serve path: pick a cloud, click install, it is running.
I build multi-cloud infrastructure and self-serve developer platforms in Go, Terraform/OpenTofu, and Kubernetes — the OpenTofu modules and Helm chart catalog behind zop.dev's self-serve provisioning on AWS, GCP, Azure, and OCI, all four at feature parity.
The throughline: if a user has to hand-write YAML or read a template, the platform is not finished. Same instinct behind Helm Doctor, which grades any chart's production-readiness A–F on a deterministic rule engine.
Parity, not exceptions
A cloud is supported when every module behaves exactly as it does on the other three.
Deterministic over clever
Rules that grade identically every run beat heuristics that are usually right.
02 — git log --career
Deployment history.
release/2026.04 — HEAD
ZopDev
Apr 2026 → present
Software Development Engineer II now
Ship provisioning and deployment features for a self-serve cloud platform — users set up compute, managed Kubernetes, and databases, connect a GitHub repository, and reach production in under 30 minutes from the UI. Collaborate with the platform team on cloud cost-optimization: scheduling idle non-production resources to power down, right-sizing recommendations from utilization data, and reclaiming orphaned resources. Also manage a multi-tenant managed hosting offering where a GitHub URL or container image becomes a live, TLS-secured endpoint in ~5 minutes.
Kubernetes
Go
Helm
Multi-Tenancy
FinOps
Cost Optimization
TLS
GitHub
release/2025.07
ZopDev
Jul 2025 → Mar 2026 · 9m
Software Development Engineer I
Delivered OCI end-to-end — networking, Kubernetes, and managed databases — at full feature parity with AWS, GCP, and Azure, and owned the OpenTofu modules and Helm charts behind self-serve provisioning. Built a standalone Go microservice that auto-tags cloud resources using ONNX models trained on resource metadata, with a human-in-the-loop feedback cycle driving continuous retraining. Provided DevOps support for a multi-tenant e-commerce platform on GCP, cutting cloud spend by ~₹80K/month.
OCI
OpenTofu
Helm
GCP
Go
ONNX
Observability
release/2024.07
ZopSmart
Jul 2024 → Jun 2025 · 1y
Software Development Engineer · Intern
Built the reusable Terraform/OpenTofu modules that let zop.dev users stand up an entire cloud account in ~30 minutes — networking, compute, databases, and services — replacing module-by-module manual setup. Authored and maintained 20+ production Helm charts, each with a values schema for single-click install.
Terraform
OpenTofu
Helm Charts
Values Schema
AWS
GCP
Azure
release/2023.10
Centillion Labs
Oct 2023 → Mar 2024 · 6m
Platform Engineer · Intern
Architected the Terraform infrastructure-as-code baseline for a fintech client across AWS and GKE, bringing dev, staging, and production live in ~1 hour, with CircleCI pipelines running automated checks on every deployment. Shipped a production AI chatbot on Vertex AI and Bedrock.
Terraform
GKE
CircleCI
AWS
Vertex AI
AWS Bedrock
release/2023.03
Jeonbuk National University
Mar 2023 → Sep 2023 · 7m
Deep Learning Engineer · Intern · Remote
Trained GAN models for facial prediction, improving image fidelity through iterative architecture tuning and systematic data preprocessing. Automated the training and code-execution workflows with shell scripting, cutting a full fine-tuning run to ~1 hour and making experiments reproducible across the research team.
Python
GANs
Bash
Reproducibility
02a — git diff --stat
4
Clouds at feature parity
20+
Production Helm charts shipped
30m
Cloud account to running services
80K
₹/month cloud spend removed
The work, diffed.
Two years of commits, squashed into numbers.
03 — cat stack.yaml
The toolchain, declared.
production toolkit
Built shipping multi-cloud provisioning and production Helm charts that other engineers deploy on every day.
1stack:2 languages: [go, python, bash, sql]3 iac: [terraform, opentofu]4 orchestration: [kubernetes, docker, helm]5 cloud: [aws, gcp, azure, oci]6 delivery: [github-actions, circleci]7 backend: [gofr, rest-apis, microservices, redis]8 observability: [grafana, prometheus, mimir, loki]9 ai-ml: [vertex-ai, bedrock, onnx]10 tooling: [linux, git, github]1112mode: ship_the_platform # self-serve • four clouds • zero hand-written YAML
Terraform, OpenTofu, AWS, GCP, Azure, OCI
Kubernetes, Helm, Docker, GitHub Actions, CircleCI
Go, GoFr, Redis, Grafana, Prometheus, Loki, Mimir
04 — git remote -v
Shipped, not claimed.
open source · maintainer
zopdev
helm-charts — 20+ production charts with values schemas for single-click install
maintainer · ★ 1.0k
open
zopdev
opentofu-modules — networking, compute, databases across AWS, GCP, Azure, OCI
maintainer · ★ 877
open
stackshy
cloudemu — simulates real AWS, Azure, and GCP resources in-memory, on your machine
contributor · ★ 142
open
project · helm doctor
live · free · nothing stored
Your Helm chart,
Your Helm chart,
examined & cured.
Grades any Helm chart A–F for production-readiness across Security, Reliability, and Upgradeability — then writes the exact YAML fix for every finding. Charts are fully rendered, exactly as Kubernetes would deploy them, then checked against 50 deterministic rules.
- 50-check rule engineDeclarative CEL expressions plus Go rules. Same chart, same values, same grade — every run, typically under 150 ms.
- Guided chart builderDescribe a workload or paste a docker-compose file, and download a chart born with a clean bill of health — live-graded as you configure.
- Values-aware gradingSupply your production values.yaml and get graded on what you would actually deploy. Helm test hooks excluded, so they never drag the grade down.
education
B.Tech · CGPA 8.64 / 10
Artificial Intelligence & Data Science
Kongu Engineering College, Erode, Tamil Nadu