India · open to new roles

Nandakumar Arumugam

Engineering Manager · Platform Engineering & AI Infrastructure · Cloud Infrastructure · Global Team Leadership

Engineering leader with 12 years in DevOps and SRE. I lead a global team of 12 across India, the US, the UK and Singapore, and own a $4M cloud estate, a 13-tool engineering platform and an AI-first developer experience. I fix root causes, so nobody gets paged at 2 AM for the same problem twice.

emailnandakumar.0413@gmail.com LinkedIn ↗ GitHub ↗
12 yrs
DevOps · SRE · Platform
$4M
Cloud estate · 25 AWS accounts
12
Engineers · 4 countries
13
Tools in my platform
Measured impact

Results, in the units they were measured in

$250K/ year

Annualized cloud savings from an enterprise FinOps program across 25 AWS accounts.

Illumina
2d → 2h

Developer turnaround after I shipped an AI-first access platform.

Illumina · 50 teams
−40%

Operational overhead after moving EC2 workloads to Amazon EKS.

Illumina
~80%

Of TeamCity workloads migrated to GitHub Actions.

Illumina · $50K licences saved
P0/P1

Incident commander for Marketplace, including Black Friday war rooms.

Walmart Global Tech
SOC2 · HIPAA

Access governance and secrets management built into regulated environments.

Illumina
Selected work

Problems I owned end to end

ILLUMINA · 2024–NOW
$250K
annualized savings

Enterprise FinOps across 25 AWS accounts

Problem
A $4M, 25-account estate needed a reliable way to find waste and get it fixed.
What I did
Built the program from scratch: idle-resource detection, rightsizing, and Jira-based remediation so every finding lands with an owner. Backed it with a real-time executive dashboard in Python, Postgres and Grafana.
Result
$250K in annualized savings, with cost tracked as a metric.
PythonPostgresGrafanaJiraAWS
ILLUMINA · 2024–NOW
2d → 2h
turnaround, 50 teams

AI-first developer experience

Problem
Identity and access requests went through a multi-step manual workflow that took about two days.
What I did
Built a conversational AI interface for identity and access workflows, so a developer asks once in plain language and the guardrails are enforced inside the flow.
Result
Turnaround fell from 2 days to 2 hours across 50 teams.
PythonLLM integrationAgentic workflowsIAM
ILLUMINA · 2024–NOW
−40%
operational overhead

CI/CD and Kubernetes modernization

Problem
Builds ran on licensed TeamCity and workloads on EC2, carrying licence cost and operational upkeep.
What I did
Migrated about 80% of TeamCity workloads to GitHub Actions and moved EC2 workloads to Amazon EKS. Now modernizing delivery with Argo CD and Argo Rollouts.
Result
$50K in licences saved and 40% less operational overhead.
GitHub ActionsEKSTerraformArgo
ILLUMINA · 2024–NOW
One view
for DORA, SLAs and KPIs

Engineering Insights Platform

Problem
Delivery and reliability data lived in separate tools, so there was no shared answer to "how are we doing?"
What I did
Architected a cross-tool platform that surfaces DORA metrics, SLA conformance and executive KPIs as a single source of truth.
Result
Engineering and leadership work from the same numbers.
DORASLA/SLOGrafanaPrometheus
WALMART · 2020–2024
Peak-ready
Black Friday Marketplace

Site reliability for Marketplace at peak scale

Problem
High-traffic Marketplace services had to stay up through peak events, with recurring failures paging engineers.
What I did
Led pre-event readiness reviews and real-time war-room coverage, served as incident commander for P0/P1 events, and built Golden Signals and SLI-vs-SLO dashboards. Designed self-healing automation with xMatters, Grafana and StackStorm, led DR drills, and moved real-time workloads from VMs to Kubernetes.
Result
Measurably lower MTTR, common failures fixed without a human in the loop, and a maturity model leadership used to plan improvements.
Incident commandStackStormSLI/SLOKubernetes
Open source

Things I build on my own time

Each project tackles a reliability or security problem I've seen in production.

Augur

Predictive anomaly detection for microservices. Connects to Prometheus, auto-discovers services, and learns each metric's normal rhythm with Prophet to flag anomalies before they become incidents.

ProphetPrometheusFastAPIReact
Sentinel

Reliability conformance dashboard. Real-time SLI, SLO and SLA tracking with AI-powered breach analysis, built on a hexagonal FastAPI backend.

SLI/SLOFastAPIReactClaude API
VeloMatrix

Engineering maturity matrix cockpit. Portfolio-level 0–10 scores with Sit / Crawl / Walk / Run bands per dimension, evidence provenance and pluggable connectors.

FastAPIReactHexagonalDocker
PySentinel

Python module security scanner. Give it a PyPI name or GitHub URL and get a SAFE / CAUTION / AVOID verdict from 14 checks, including secrets in git history, SAST and CVE scans.

gitleakssemgrepbanditpip-audit
How I work

Operating principles

I can't fix what I can't see

Every system I build ships with service level indicators. Observability comes first, not after the first outage.

Root causes over patches

I dig until the real cause is found, then log the tech debt the same day so it doesn't come back.

Security from design, not after

Scanning, secrets management and compliance controls are built into the pipeline from the start, so audits are routine.

API-first, automate the second time

If a task happens twice, it becomes a tool. Self-service beats a ticket queue.

Career · to scale

Twelve years, five teams

  1. Jun 2024 – now

    DevOps Engineering Manager · Illumina

    Lead a global DevOps and SRE team of 12 across India, US, UK and Singapore. Own a 13-tool engineering platform, $4M in annual spend and the full people lifecycle. FinOps, CI/CD and EKS modernization, the Engineering Insights Platform, the AI-first developer experience, and CrowdStrike SHRA integrated with JFrog Artifactory for automated vulnerability assessment.

    AWS · EKS · Terraform · Ansible · GitHub Actions · Python · Vault · Grafana · Prometheus · Loki · Splunk · Cribl · CrowdStrike · JFrog · SonarQube · Mend · PagerDuty

  2. Sep 2020 – May 2024

    Sr. SRE & DevOps Engineer · Walmart Global Tech

    Site reliability for Marketplace through peak events. Readiness reviews, war rooms, incident command, Golden Signals dashboards, DR drills and runbooks, VM to Kubernetes migration, self-healing automation, and an engineering maturity framework.

  3. Sep 2018 – Sep 2020

    Tech Lead (Senior DevOps Engineer) · Cognizant

    Infrastructure and deployment automation with Terraform, Jenkins, Ansible, Liquibase and Python. Jenkins pipelines spanning build, deploy, smoke testing and promotion. Defined the GitLab branching model and code-promotion workflow, and shifted security left with DAST/SAST.

  4. May 2017 – Sep 2018

    Senior DevOps Engineer · Aspire Systems

    Designed a Client Assembly Factory to standardize client creation and dependency injection, improving scalability and reducing onboarding friction.

  5. Jul 2014 – May 2017

    DevOps Engineer · Tata Consultancy Services

    Jenkins CI/CD for Java applications and automated database deployments, improving release efficiency and reducing deployment errors.

Core skills

What I build with

Cloud & infraAWS, Azure
Automation & IaCTerraform, Ansible, Python, Shell, Liquibase
CI/CD & DevSecOpsGitHub Actions, Jenkins, Azure DevOps, TeamCity, SonarQube, Mend
Observability & SREDynatrace, Grafana, Prometheus, Loki, Tempo, Splunk, CloudWatch, PagerDuty, StackStorm, Graylog
SecurityHCP Vault, CrowdStrike, CIS Benchmarks, SOC2, HIPAA
LeadershipGlobal team management, hiring and performance, platform strategy, incident and DR leadership, cost governance
Data & messagingMySQL, Oracle, Postgres, Kafka, RabbitMQ, Hazelcast
Web & edgeJBoss, Tomcat, Nginx, Akamai, F5, Cloudflare, WAF
Other toolsJira, Confluence, Artifactory, ELK
EducationB.Tech in Mechatronics, SASTRA University (2010–2014)
Get in touch

Let's talk about your platform

I'm looking for engineering leadership roles in platform engineering, SRE and AI infrastructure. I'm based in India and open to relocation.