DevOps · SRE · Platform Engineering

I build infrastructure that's reliable, observable, automated, secure and cost-aware.

I run production infrastructure on AWS, and since 2023 I have owned DevOps end to end for an ad and deals platform. I build the pipelines, environments and guardrails that let a product team ship daily without waking anyone up.

Email meRésumé LinkedIn GitHub
What I work on: AWS, CI/CD, reliability and observability. Each links to the matching skills.

Naren Anandan

DevOps & Platform Engineer

in production since 2021

AWSElastic Beanstalk · EC2 · CloudFront · CloudFormationCI/CDGitHub Actions · runners · build cacheReliabilityIncidents · capacity · SLOsObservabilityNew Relic · CloudWatch · OpenTelemetry

How I work

  1. Boring is a feature

    Predictable deploys and quiet pagers beat clever systems nobody can debug at 3 a.m.

  2. Everything as code, reviewed

    Infrastructure, pipelines and docs live in git, change through pull requests, and can be rebuilt.

  3. Measure before changing

    Root cause from logs, traces and latency data first. Then the smallest fix that holds.

  4. Automate the second time

    Do it by hand once and write it down. The second time it becomes a script, a check or a golden path.

Selected work

Open a card for the situation, what I did and the stack. Internal names and absolute figures are left out on purpose.

CI/CD platformMoved the company to GitHub, then made ~25 repositories build the same way30% lower licensing and operating cost, pipelines about 40% faster, and one change now rolls out across every repository.

Situation

Builds were split across Bitbucket Pipelines and one-off workflows. Every repository did CI slightly differently, and upgrades meant editing each one by hand.

What I did

Made the case for GitHub and led the migration. Then built shared actions and a standard workflow layout, pinned every action to an exact release, added a remote Gradle build cache with parallel tests, and automated rebuilds of the self-hosted runner images before they age out.

  • GitHub Actions
  • Self-hosted runners
  • Gradle
  • Bash
Data platformMoved live Elasticsearch clusters up to current 8.xClusters on a supported release, the old ones torn down, and a repeatable procedure for the next upgrade.

Situation

Search and deals data sat on an older Elasticsearch major version, with index mappings that could drift between environments.

What I did

Planned and ran the major-version migrations, each with a written runbook. Moved the clusters to Graviton, added a disk health check that adds data nodes on its own, and built a mapping check that catches drift before production.

  • Elasticsearch 8
  • EC2 Graviton
  • ILM
  • Bash
ReliabilityRoot-caused 502s and host crashes on Elastic BeanstalkScale-in no longer drops requests, and the fixes are standard environment configuration kept in git.

Situation

A busy service threw bursts of 502s and flapped between scaling up and down. Another host fell over under memory pressure.

What I did

Traced the 502s to a connection-drain window shorter than the slowest requests, and the crashes to runtime memory limits that added up to more than the container had. Switched the load balancer to least-outstanding-requests routing and set drain and memory limits from measured latency and usage.

  • Elastic Beanstalk
  • ALB
  • Auto Scaling
  • New Relic
Release engineeringVersioned delivery and one-step rollback for the web appsEach app ships on its own version, and rollback was proven in staged production runs before go-live.

Situation

Front-end releases were all versioned as patches, and getting back from a bad deploy meant redeploying by hand.

What I did

Built a CloudFront and S3 delivery platform for the Angular apps with per-app release scoping, versions driven by commit messages, blue/green distributions, and source maps wired into error tracking.

  • CloudFront
  • CloudFormation
  • S3
  • Nx
Developer platformA one-command dev environment, and a safe sandbox for AI coding agentsNew developers start from one script, and agents can read, draft and review infrastructure changes while a person still applies them.

Situation

Setting up a working environment took days, and the team wanted AI coding agents without handing them production access.

What I did

Shipped versioned bootstrap scripts for Windows and Ubuntu with a local AWS emulator in Docker Compose. Designed an isolated sandbox for agents, kept apart from developer machines and production, and moved team docs to a docs-as-code site.

  • Docker Compose
  • PowerShell
  • CloudFormation
  • IAM
  • MkDocs

Experience

5+ years of continuous infrastructure work.

  1. Mar 2023 — Present

    DevOps Platform Engineer · ShopLiftr

    DevOps for an AWS platform of Groovy/Spring Boot services, Node and Angular apps, and Elasticsearch.

    • Run the AWS estate (Elastic Beanstalk, EC2, CloudFormation, CloudFront, S3, Lambda) at 99.99% uptime, and cut monthly spend by 28% through rightsizing.
    • Led the move from Bitbucket to GitHub: 30% lower licensing and operating cost, and pipelines 40% faster.
    • Took releases from weekly to daily with GitHub Actions, and cut deployment errors by 90%.
    • Containerised every service with Docker and Docker Compose: deployment time down 70%, and local environments that match production.
    • Built monitoring and alerting on CloudWatch and New Relic, cutting mean time to resolve by 45%.

  2. May 2022 — Feb 2023

    DevOps Specialist · Pythian

    Cloud consultancy, mainly on Google Cloud, across client and internal teams.

    • Split a legacy monolith into 10+ containerised services, making releases 50% faster.
    • Automated CI/CD with Jenkins and GitLab CI: 75% less manual deployment work, release cycles from 3 days to under 8 hours.
    • Redesigned client environments on GCE, GKE, Cloud SQL and Cloud Build for 99.95% availability.
    • Added dependency scanning and secret detection to pipelines, and observability on Prometheus, Grafana and Stackdriver that cut time to detect by 40%.

  3. Oct 2021 — Apr 2022

    Network Operations Center Engineer · Eastlink

    24×7 operations for a major telecommunications provider.

    • Monitored 500+ customer networks, servers and dedicated links, holding a first-response SLA of 95% or better.
    • Resolved 85% of incidents without escalation, ranked in the top 10% of the team, and trained 5+ new engineers.

Before that: Graduate Teaching Assistant at Dalhousie University (2019–2021). I helped move a 249-student engineering course online during the pandemic.

Skills

In production at work, every weekHands-on experience

Cloud & data

  • AWS (in production weekly)
  • Elastic Beanstalk (in production weekly)
  • CloudFront (in production weekly)
  • Elasticsearch 8 (in production weekly)
  • EKS / ECS / Fargate (hands-on)
  • Google Cloud · GKE (hands-on)
  • Azure (hands-on)
  • Redis (hands-on)

AWS daily; Google Cloud and Azure from consulting work.

Infrastructure as code

  • CloudFormation (in production weekly)
  • Terraform (hands-on)
  • AWS CDK (hands-on)
  • Crossplane (hands-on)
  • Pulumi (hands-on)
  • Ansible (hands-on)

Reviewed, versioned templates, with change sets before anything touches production.

Kubernetes & GitOps

  • Kubernetes (hands-on)
  • Helm (hands-on)
  • Argo CD (hands-on)
  • Flux (hands-on)
  • Karpenter / KEDA (hands-on)
  • Docker (in production weekly)

Cluster delivery driven from git, with autoscaling sized from real load.

CI/CD & release

  • GitHub Actions (in production weekly)
  • Self-hosted runners (in production weekly)
  • Gradle build cache (in production weekly)
  • Blue/green & rollback (in production weekly)
  • Jenkins (hands-on)
  • GitLab CI (hands-on)

Shared workflows, fast feedback, versioned releases and a tested way back.

Observability

  • New Relic (in production weekly)
  • CloudWatch (in production weekly)
  • OpenTelemetry (hands-on)
  • Prometheus (hands-on)
  • Grafana (hands-on)
  • ELK (hands-on)

Alerts tied to real failure modes, not dashboards nobody opens.

Reliability (SRE)

  • Incident response (in production weekly)
  • Blameless postmortems (in production weekly)
  • Capacity & autoscaling (in production weekly)
  • Runbooks (in production weekly)
  • SLOs & error budgets (hands-on)
  • Load & chaos testing (hands-on)

Evidence-first root cause, written up so it does not happen twice.

Security & supply chain

  • IAM least privilege (in production weekly)
  • Pinned, scanned CI (in production weekly)
  • Secrets Manager / Vault (hands-on)
  • Policy as code (OPA, Kyverno, Checkov) (hands-on)
  • SBOM & signing (Syft, cosign) (hands-on)
  • GuardDuty · Security Hub · WAF (hands-on)

Least privilege by default, and dependencies pinned to exact versions.

Platform & cost

  • Dev environments (in production weekly)
  • Docs as code (in production weekly)
  • AWS cost & rightsizing (in production weekly)
  • Backstage (hands-on)
  • Kubecost (hands-on)

Golden paths for developers, and spend that is measured, not guessed.

Code & AI-assisted ops

  • Bash (in production weekly)
  • PowerShell (in production weekly)
  • Groovy (in production weekly)
  • Claude Code & MCP (in production weekly)
  • Python (hands-on)
  • TypeScript (hands-on)

Scripts and tools that remove manual steps; agents that read, draft and review.

Credentials

Every certification opens its issuer's verification page.

Google Cloud Professional Cloud DevOps Engineerverify ↗HashiCorp Certified: Terraform Associateverify ↗Google Cloud Associate Cloud Engineerverify ↗Microsoft Certified: Azure Developer Associateverify ↗
MEng, Electrical & Computer EngineeringDalhousie University, Halifax2018–2021
BE, Electronics & Electrical EngineeringAnna University, Chennai2014–2018

Let's talk.

Email is the fastest way to reach me. I usually reply within a day.