Shubham Kumar - Platform Engineer

About me

Platform Engineer at Dremio, managing multi-client Dremio deployments across Kubernetes flavors - EKS, OpenShift, and on-prem. Before that, 4+ years at Modak Analytics building Big Data infrastructure on AWS and Cloudera CDP. My focus is Kubernetes-native architecture, observability, and automation - building systems that are reliable, cost-efficient, and don't surprise you at 3am.

Current role

Platform Engineer @ Dremio

Location

Hyderabad, India

Open to

Freelance & consulting

Skills

Cloud

AWS

GCP

Azure

Kubernetes

Helm

Docker

ArgoCDKarpenter

Observability

Prometheus

Grafana

Datadog

Node Exporter

Grafana Alloy

Loki

Promtail

IaC & CI/CD

Terraform

Jenkins

GH Actions

Bitbucket

GitLab

Version Control

Git

GitHub

Bitbucket

GitLab

AWS Services

EKS

S3

Lambda

IAM/IRSA

ALB/NLB

SQS

SES

Route53

ECR

EC2

VPC

Security

K8s RBAC

KeycloakAppArmorSeccomp

OPA/Gatekeeper

FalcoAdmission ControllersNetwork Policy

Cilium

Trivy

Big Data

Airflow

Spark

HadoopDremio

Languages

Python

Shell

PowerShell

Work experience

Platform Engineer / Dremio Administrator

Dremio · Hyderabad, India

Dec 2025 - Present

Managing multi-client Dremio deployments on Kubernetes across EKS, OpenShift, and on-prem clusters.
Observability across client environments using Datadog, Grafana, and Prometheus.
OAuth/SSO integrations using Keycloak across client deployments.
Storage integrations across AWS S3, GCP, and Scality Artesca.
Helm chart authoring and management for Dremio and supporting services.
Built the Dremio Performance Dashboard (DPD) - Flask-based K8s app for query performance visibility, published as a Helm chart to GitHub.

Platform DevOps Engineer

Modak Analytics LLP · Hyderabad, India

Jun 2021 - Dec 2025

Deployed and managed Big Data platforms across AWS and Cloudera CDP.
End-to-end observability with Prometheus, Grafana, Alertmanager - 50% reduction in manual monitoring.
CI/CD pipelines via Jenkins and ArgoCD - 30% faster deployments.
Managed 5+ AWS EKS clusters with custom CNI networking and EBS/EFS storage configurations.
Deployed Apache Airflow and Marquez on EKS via Helm; integrated OpenLineage for data lineage.
Migrated ML workloads to Spark on EKS - eliminated vendor instances, 40% cost reduction.
Configured Karpenter for intelligent node auto-scaling and cost optimization.
Managed Neo4j end-to-end: backup, monitoring, upgrades, and licensing cost reduction.

Projects

EKS Observability Stack

Centralized logging and monitoring with Grafana, Loki, Prometheus, and Alloy across 5+ EKS clusters.

PrometheusGrafanaLokiEKS

Automated Alert-to-Ticket Pipeline for Support

Built an event-driven pipeline where Alertmanager routes qualifying alerts through SQS to a Lambda function. The Lambda queries Dremio Cloud to map the alerting cluster to a customer, checks whether that customer has certain entitlement, and based on alert severity plus that entitlement check, it automatically opens a high or mid-priority Zendesk ticket.

AWS LambdaSQSSecrets ManagerAlertmanagerDremio CloudZendeskOIDC & IRSA