Cloud & Kubernetes Platforms
Designing and operating AWS environments (EKS, EC2, S3, Lambda, VPC) and GKE clusters — autoscaling, capacity planning, and cost optimization for production workloads.
- AWS
- EKS / GKE
- Terraform
- Helm
$ whoami
devops engineer · platform builder · reliability nerd
$ uptime --mindset
high availability, least privilege, automate everything
$
I'm Harshit — a systems-focused engineer. I build infrastructure that just works: fast, secure, observable, and boring in the best possible way.
I own the full path to production — networking, Kubernetes, CI/CD, and every guardrail around it. In the last two years I've led a Kubernetes migration, moved CI/CD from Jenkins to GitHub Actions with GitOps delivery, migrated configuration management to Ansible, and driven an enterprise Bitbucket-to-GitHub move — with security and IAM designed in, not bolted on.
Lately my scope has widened: testing infrastructure that makes every release safer, infrastructure for AI workloads and AI-assisted developer tooling, and treating security, compliance and governance as first-class platform features.
Designing and operating AWS environments (EKS, EC2, S3, Lambda, VPC) and GKE clusters — autoscaling, capacity planning, and cost optimization for production workloads.
Building and migrating delivery pipelines — Jenkins to GitHub Actions, GitOps with ArgoCD and FluxCD — so shipping to production is boring, fast, and reversible.
IAM design and least-privilege access, secrets management, vulnerability remediation, audit readiness, and access-governance processes that hold up under review.
Monitors, dashboards, alerting and on-call workflows with Datadog, Grafana, CloudWatch and Zenduty — catching issues before customers do and shortening MTTR.
Standing up infrastructure for AI workloads and AI-assisted developer tooling, plus test environments and automation that make every release safer to ship.
VPC design, DNS, TLS, reverse proxies, load balancers and firewalls — with the Linux administration depth to debug anything from packet to process.
Bangalore, India
Bangalore, India
Bangalore, India
Selected platform initiatives I've led or built, told as problem → approach → outcome. Details are intentionally generic — the patterns are what matter.
ProblemLegacy VM-based workloads were expensive and slow to scale.
BuiltA phased migration to Kubernetes (EKS): workload audit, right-sizing, autoscaling, and a cutover with rollback paths at every step.
Result40% lower infra cost, faster deployments, self-healing workloads.
ProblemA sprawling Jenkins estate: high maintenance, snowflake pipelines.
BuiltDelivery rebuilt on GitHub Actions — reusable workflows, OIDC cloud auth with no long-lived keys, ArgoCD for GitOps deployment.
ResultPipelines as reviewable code, secretless auth, consistent releases.
ProblemServer configuration lived in scripts and tribal knowledge.
BuiltEverything moved into Ansible — inventories, roles, idempotent playbooks — so any environment can be rebuilt or audited from source control.
ResultZero drift — config changes reviewed like code.
ProblemAn organization's entire source control needed to move — without losing history or trust.
BuiltFull repository migration with history, branch protections, team/permission mapping, and IAM aligned to security standards from day one.
ResultZero-loss migration with stronger access governance than before.
ProblemReleases needed a safety net beyond unit tests.
BuiltTest infrastructure and automated environments that exercise changes realistically before they reach production.
ResultRegressions caught before customers see them.
ProblemTeams wanted AI capability; the platform had no safe way to provide it.
BuiltProvisioning for AI workloads and rollout of AI-assisted developer tooling — with access control, cost visibility and usage governance from the start.
ResultAI adoption with governance built in, not bolted on.
ProblemSecurity treated as an afterthought becomes audit pain and real risk.
BuiltLeast-privilege IAM across cloud and tooling, secrets management, vulnerability remediation cycles, security monitoring, and documentation that keeps audits calm.
ResultA platform that is compliant by construction — and provably so.
Go backend API that executes OS commands with a real-time frontend for input and output — a small remote-ops console.
Hands-on Kubernetes infrastructure lab — manifests, shell tooling and cluster experiments used for real platform learning.
Google Kubernetes Engine deep-dive — multi-cloud K8s beyond EKS: cluster setup, workloads and GCP-native patterns.
GitOps monorepo wired to FluxCD — declarative delivery where a git push is the deployment.
Smart India Hackathon project — image-based authentication tested to resist brute-force and dictionary attacks.
Infrastructure-as-Code workbench — HCL modules and AWS provisioning patterns built from scratch.
I write plain-language deep-dives on infrastructure at harshitpaneri.hashnode.dev.
Salesforce platform fundamentals
Microsoft Technology Associate
Machine learning program by Amazon
MongoDB University
JavaScript fundamentals certification
Internship certification program
Spoke at the Google I/O event hosted by GDG Udaipur.
Worked with cloud-native open-source projects including Meshery (CNCF) and SigNoz; ran a 10★ Hacktoberfest starter repo for first-time contributors.
Mentor at Social of Code '23 and Script Winter of Code '22; DevRel mentee at Keploy 4.0.
Organizer and active participant across Google Developer Student Clubs and FOSS events.
Whether it's an infrastructure problem worth solving, a role worth discussing, or just a good conversation about systems — my inbox is open.
Say hello