KRKuldeep Singh Rathore
Open to remote roles, contracts & UAE

DevOps Lead · Platform & MLOps

Kuldeep Singh Rathore

I turn fragile infrastructure into platforms teams can trust: Kubernetes and event streaming at scale, MLOps for GPU workloads, and cloud cost that stays under control.

Download résumé Get in touch LinkedIn
7+ years since 2019AWS · Azure · GCPBased in New DelhiNotice 1 month
What I built for EY DigiGST 2.0simplified architecture
GST filing trafficEY clients
Burst-heavy around filing deadlines
HTTPS
Azure Kubernetes Service
AMS messaging serviceHPA · autoscaled
Right-sized requests so new pods take traffic
produce · TLS + JWT
Apache PulsarHelm
Brokersrouting
Bookiesstorage
ZooKeeperStatefulSet
consume
Consumer APIHPA
Pod scheduling fixed
Pulsar IO sinks
Automated sink config
cert-manager TLSJWT authGrafana node metricsdedicated node pools (next)
Results and what I changed are in the case study.Read the case study →
7+ yrsDevOps, cloud and MLOps across 3 clouds
−80%Manual work in ML deployment (LifeSafety.AI)
+25%Faster GPU model training
99.9%Uptime held on Yondr's DCIM platform
500+Vulnerabilities fixed for SOC 2

About

Infrastructure is a product. I build it like one.

I'm a DevOps lead with 7+ years of turning fragile infrastructure into platforms teams can trust. Today I lead DevOps at CoreOps.ai, where I run production platforms for enterprise clients like EY, Panasonic and NM India, own our multi-cloud cost governance, and sit on the pre-sales team shaping RFPs and architectures before a line of code is written.

My sweet spot is where infrastructure meets AI: Kubernetes and event streaming at scale, MLOps pipelines for GPU workloads, and the FinOps work that keeps all of it affordable. I've also been the person on call when things break, including leading incident response on a compromised UAT environment.

I'm the technical lead for CoreOps.ai's AWS, Red Hat and SUSE partnerships, and I'm certified across Kubernetes (CKA), AWS and Anthropic's Claude.

Current roleDevOps Team Lead, CoreOps.ai
CloudsAWS · Azure · GCP
FocusKubernetes, MLOps, FinOps
ClientsEY · Panasonic · NM India · Yondr
Looking forRemote · Contract · UAE
Notice period1 month

Case studies

Three platforms, start to finish

What the problem was, what I did, and what changed.

EY · DigiGST 2.0 · Azure

A GST messaging platform tuned to 21× its expected peak

EY's GST compliance platform has to survive filing-deadline bursts. I designed the Apache Pulsar platform on AKS and tuned it through EY's performance testing.

  • Deployed Pulsar with Helm: ZooKeeper StatefulSets, cert-manager TLS, JWT auth and automated sinks.
  • Right-sized CPU and memory requests so autoscaled pods actually took traffic.
  • Traced node CPU saturation in Grafana to Pulsar and the app sharing one node pool.
  • Proposed dedicated AKS node pools as the path to 5,000 TPS.
Error-free throughput, TPSEY load tests
Expected peak
37.75
Before tuning
~450
After tuning
820
100%0%Consumer API failures
21.7×Headroom over expected peak
Next step: 5,000 TPS with dedicated node pools for Pulsar and the messaging service.
CoreOps.ai · internal platform

AgentCORE: an MLOps platform for AI agents

CoreOps.ai needed one place to build, train, deploy and monitor AI agents. I designed, built and launched it as the sole DevOps engineer.

  • Separate staging, UAT and production environments with promotion between them.
  • Jenkins CI/CD with SonarQube quality gates on every change.
  • Prometheus and Grafana monitoring for models, services and infrastructure.
  • Rightsized the production VM as part of company-wide cost governance.
Agent lifecycle on AgentCOREsimplified
1
Build
Code and agent configs, SonarQube quality gate
2
Train
Model and agent training jobs
3
Deploy
Jenkins pipelines promote through environments
stagingUATproduction
4
Monitor
Prometheus and Grafana dashboards and alerts
Podtech · LifeSafety.AI

Taking computer-vision safety models into production

LifeSafety.AI's models worked in development but deploying them was manual and slow. I architected the MLOps infrastructure that took them to production.

  • Automated the image-processing pipeline and model deployment with Jenkins.
  • Managed and tuned NVIDIA multi-GPU training servers.
  • Built real-time Prometheus and Grafana dashboards, working daily with data scientists.
Model path to productionsimplified
1
Image processing
Automated pipeline replaces manual prep
2
GPU training
NVIDIA multi-GPU servers, tuned for throughput
3
Deployment
Jenkins ships models to production
4
Observability
Model, pipeline and infrastructure health
−80%Manual intervention
+25%Training speed

Panasonic ERP (AX) modernisation CoreOps.ai

Mirror servers, NBFC partner integrations, database parity and an Azure Data Factory migration in overnight change windows.

FinOps across AWS, Azure and GCP CoreOps.ai

Two-account AWS FinOps assessment for Panasonic; guardrails, rightsizing and a billing waiver for CoreOps.ai's own spend.

Yondr DCIM: CI/CD and containers Podtech

Bitbucket and Jenkins to Azure DevOps (40% faster pipelines), Docker and Kubernetes, PRTG and Grafana monitoring.

Incident response on a compromised server CoreOps.ai

Forensic root-cause analysis, rebuild from a hardened production clone, and vendor remediation.

Pre-sales & partnerships

The technical side of how CoreOps.ai wins work

AWS

Technical lead

Partner Central opportunities, landing-zone designs and partner-tier requirements.

Red Hat

Technical lead

OpenShift and OpenShift Virtualization deal registrations.

SUSE

Technical lead

Partner onboarding and tier planning.

RFPs & bids

Pre-sales

Cloud sizing, solution architecture and technical compliance for government and enterprise bids.

Experience

7+ years, three companies

Dec 2024 – Present · Noida

DevOps Team Lead · Senior DevOps & MLOps Engineer

CoreOps.ai
  • Lead the DevOps team: assign and review work, report weekly to leadership, interview and hire.
  • Production platforms for EY (DigiGST 2.0), Panasonic (AX modernisation) and NM India (N-Gage hosting, Phase-1 data foundation proposal).
  • Built and launched AgentCORE; govern costs across AWS, Azure and GCP.
  • Technical lead for AWS, Red Hat and SUSE partnerships; pre-sales, RFPs and solution architecture.
  • Microsoft 365 administration and tenant migration.
May 2022 – Nov 2024 · Gurgaon

Platform DevOps & MLOps Engineer

Podtech
  • LifeSafety.AI: end-to-end MLOps for computer-vision models.
  • Yondr DCIM: Azure DevOps migration, Docker and Kubernetes (deploys 50% faster).
  • Monitoring that cut downtime 30%; SonarQube gates cut technical debt 20%.
Jun 2019 – May 2022 · Jaipur

DevOps & Cloud Engineer (Linux / Cloud Admin)

Global IT Providers
  • Led the SOC 2 effort by fixing 500+ security vulnerabilities.
  • Ran AWS infrastructure: EC2, ELB, S3, IAM, Route 53, WAF, KMS, RDS, VPC.
  • Automated deployments, backups and log rotation with Bash and Python.
2012 – 2016

B.Tech, Electronics & Communication Engineering

Rajasthan Technical University

Stack

Tools I use in production

Cloud

AWSAzureGCPCloudflareBedrock

Containers

KubernetesAKSHelmDockerHPAStatefulSets

CI/CD

JenkinsAzure DevOpsGitHubBitbucketSonarQubeTerraform

MLOps & AI

Model pipelinesNVIDIA multi-GPUClaudeLLM platforms

Streaming & data

Apache PulsarMySQLMongoDBSQL ServerData Factory

Observability

PrometheusGrafanaCloudWatchSumo LogicPRTGKibana

Security

SOC 2Incident responseCyberArkFortiGateTLS · JWT

FinOps & platforms

RightsizingSavings PlansBudgetsMicrosoft 365Linux

Certifications

Certified and current

CKA

Certified Kubernetes Administrator

The Linux Foundation · CNCF

Valid until Jan 2027
AWS

AWS Certified Solutions Architect – Associate

Amazon Web Services

Valid until May 2028
AI

Claude Certified Architect – Foundations

Anthropic

Valid until Aug 2027
APN

AWS Partner: Technical Accredited

AWS Partner Network · 2026

RH

Red Hat Partner Program

Ready tier · 2026

B.T

B.Tech, Electronics & Communication

Rajasthan Technical University · 2016

Contact

Open to remote roles and contract work

DevOps, platform and MLOps roles: full-time remote, contracts, or relocation to the UAE. One-month notice.

Remote, India or international teams
Contracts and consulting engagements
On-site or hybrid in Dubai / Abu Dhabi
Emailrathorethechamp@gmail.com LinkedInin/kuldeep-devops GitHubkuldeep-devops01
Résumé (PDF)