Buffalo, New York · U.S. work authorized
I build cloud infrastructure
that stays up.
I build and operate AWS infrastructure, Kubernetes platforms, and CI/CD systems that stay up in production. I work across AWS, Amazon EKS, Terraform, and GitOps delivery, and I care about cost, recovery time, and leaving a clean trail behind me.
- AWS
- Kubernetes
- Amazon EKS
- Docker
- Terraform
- GitLab CI
- GitHub Actions
- ArgoCD
- Linux
Hands-on with the systems that need to stay up
Seven years spent where uptime, cost, and recovery time are things people actually notice.
AWS SAA · CKA
Certified
I am a Cloud DevOps Engineer with 7 years of experience building and running AWS infrastructure, Kubernetes platforms, and CI/CD systems in production. Most of that time has been spent close to real workloads, where uptime, cost, and recovery time are things people actually notice.
Right now I work at VistaJet, operating and supporting global aviation workloads on Amazon EKS. I owned the cutover plan for a zero-downtime EKS migration, cut AWS spend by 18 percent, and brought mean time to recovery down by 40 percent by pulling monitoring and alerting into one place.
I focus on the infrastructure work that usually decides whether systems stay reliable: reusable Terraform modules, GitOps workflows, practical runbooks, and observability that helps teams recover faster. When I want to go deeper on the fundamentals, I build them from scratch in a bare-metal Kubernetes homelab.
Cloud infrastructure
AWS across compute, storage, networking, and IAM, with Terraform describing it so it can be reviewed, versioned, and rebuilt.
Kubernetes platforms
Operating workloads on Amazon EKS at work and building a cluster by hand with kubeadm at home to understand every layer.
Delivery and GitOps
CI/CD on GitLab and GitHub Actions, with ArgoCD delivering Helm releases through dev, staging, and production.
Reliability
Observability on Prometheus, Grafana, and Loki, paged through PagerDuty, with runbooks that shorten recovery.
7 yrs
Cloud & DevOps experience
18%
AWS spend reduced at VistaJet
40%
Mean time to recovery reduced
2
Certifications (AWS SAA, CKA)
Certifications: AWS Certified Solutions Architect - Associate (SAA-C03), Certified Kubernetes Administrator (CKA). Every figure above comes straight from my resume.
A day in my Cloud DevOps work
What the job actually looks like. Less about the tool list, more about keeping production healthy and unblocking the people who ship to it.
Check overnight alerts and the on-call handoff
Start of dayFirst look is anything that paged or drifted overnight. I confirm nothing is still degraded before moving on.
- PagerDuty
- Grafana
- Prometheus
Read the dashboards and logs
MorningA quick pass across metrics, traces, and logs to catch slow trends before they become incidents.
- Grafana
- Datadog
- Jaeger
- Loki
Work through team requests
Mid-morningRequests come in through tickets and chat: access, a deployment, a broken pipeline, or help debugging something. I unblock developers first.
- Jira
- Slack
- Teams
Infrastructure and Kubernetes work
MiddayTerraform and AWS changes go through a merge request and a CI plan. On the cluster side I troubleshoot workloads and keep EKS healthy.
- Terraform
- AWS
- Amazon EKS
- kubectl
Keep delivery green
AfternoonSupport the CI/CD pipelines and GitOps flow so changes ship cleanly, and help developers get deployments across the line.
- GitLab CI
- GitHub Actions
- Argo CD
- Helm
Watch production releases
Release windowFollow the golden signals during a release, ready to roll back the moment something looks off rather than pushing through it.
- Grafana
- Prometheus
- Argo CD
Document and hand off
End of dayUpdate runbooks and notes, close out the day's work, and leave a clean handoff so the next person is not guessing.
- Runbooks
- Git
The stack, and how I actually use it
Hover, tap, or tab through any tool to see how it shows up in my work. Nothing here is decorative.
Cloud
Where the workloads actually run.Containers & Orchestration
Packaging and scheduling.CI/CD & GitOps
How code reaches production.Infrastructure as Code
Infrastructure you can review.Observability
Knowing what is happening.Security & Platform
Guardrails and access.AWS
Primary cloud across EKS, EC2, S3, RDS, Lambda, VPC, IAM, Route 53, CloudWatch, ALB/NLB, and Secrets Manager.
Descriptions map to specific roles and projects on this page.
Where I have done the work
Three roles, one throughline: making infrastructure repeatable, observable, and cheaper to run.
Cloud DevOps Engineer · VistaJet
Fort Lauderdale, FL (Remote)
Oct 2022 - Present
- Owned the cutover plan for a zero-downtime migration of production workloads from a self-managed Kubernetes platform to Amazon EKS, using blue-green deployments and rollback validation, with no customer-facing downtime.
- Cut AWS spend 18 percent across EC2 and EBS by right-sizing instances, cleaning up orphaned volumes, and rolling out tag-based resource governance.
- Reduced mean time to recovery 40 percent by unifying observability on CloudWatch, Prometheus, Grafana, and Loki with PagerDuty alerting.
- Built GitLab CI/CD pipelines covering automated testing, container builds, Trivy scanning, and progressive deploys to dev, staging, and production.
- Implemented ArgoCD GitOps workflows that deliver Helm releases across environments with version-controlled, auditable change history.
- Built reusable Terraform modules for EKS, VPC, IAM, RDS, and Secrets Manager, with remote state in S3 and DynamoDB and CI-based plan workflows.
- Supported RDS upgrades and disaster recovery for business-critical databases, focused on cutover planning, replication checks, and rollback readiness.
- Partnered with security to remediate Wiz findings and supported SOC 2 audit evidence collection, and authored on-call runbooks and standardized Helm charts.
DevOps Engineer · HCHK Technologies, Inc.
New York, NY
Oct 2021 - Sep 2022
- Cut deployment time 60 percent by building CI/CD pipelines on Docker, Jenkins, and GitHub Actions.
- Automated configuration across 50+ Linux servers using Ansible Tower with role-based playbooks.
- Built and supported AWS infrastructure for data warehouse workloads across VPCs, NAT gateways, and load balancers, contributing to 99.99 percent application uptime.
- Built reusable CloudFormation templates to standardize network and compute provisioning across environments.
- Set up centralized logging and alerting with the ELK Stack integrated to CloudWatch.
- Deployed OpenVPN with LDAP integration for secure remote access and automated S3 snapshot lifecycle for backups.
DevOps Engineer · Chrysler Motors Corporation
Auburn Hills, MI
2018 - Sep 2021
- Managed 200+ AWS servers across development, staging, and production environments.
- Migrated 300+ legacy Jenkins jobs to pipeline-as-code, improving auditability and team handoff.
- Reduced manual operational work 30 percent through Python and Bash automation of routine tasks.
- Maintained CI/CD on Jenkins, Bamboo, Bitbucket, and Nexus, with Chef cookbooks for Tomcat and JBoss provisioning.
- Built optimized Docker images that shrank build times and deployment payloads.
- Contributed to AWS migration and Azure hybrid cloud connectivity, with monitoring on Splunk and Nagios.
Production problems I am comfortable investigating
Common operational patterns and how I approach them. This is how I think when something breaks, not a claim that I hit every one of these in production.
CrashLoopBackOff
- Signal. A pod keeps restarting and never goes ready.
- I check. Container logs and events, the exit code, and any recent image or config change.
- Likely path. Usually a bad config, a failing probe, or a missing dependency. Fix it or roll back the release.
- kubectl
- Loki
- Grafana
OOMKilled
- Signal. A container dies with exit code 137 and memory sits at its limit.
- I check. Memory requests and limits, recent traffic, and whether usage grows over time.
- Likely path. Right-size the limits, or fix the leak if memory only climbs.
- kubectl
- Prometheus
- Grafana
CI/CD pipeline failure
- Signal. A red pipeline is blocking a merge or a deploy.
- I check. The failing stage logs, whether it is test, build, or scan, and what changed last.
- Likely path. Fix the failing step or the flaky dependency and get the pipeline green again.
- GitLab CI
- GitHub Actions
- Trivy
High API latency
- Signal. p95 latency and error rate climb together.
- I check. Traces for the slow hop, database and downstream calls, and resource saturation.
- Likely path. Find the bottleneck, a slow query, a downstream service, or scaling, and address it.
- Jaeger
- Prometheus
- Datadog
AWS access denied
- Signal. An API call returns AccessDenied or a 403.
- I check. The IAM policy, the role trust, and the exact action and resource in the error.
- Likely path. Grant the least privilege that actually works, no broad wildcards.
- IAM
- CloudWatch
- CloudTrail
Argo CD sync failure
- Signal. An app is stuck OutOfSync or Degraded.
- I check. The diff between Git and the cluster, sync hooks, RBAC, and the image tag.
- Likely path. Reconcile the manifest and fix the drift, or roll back the offending commit.
- Argo CD
- kubectl
- Git
Failed deployment
- Signal. A rollout stalls and new pods never become ready.
- I check. Events, readiness probes, resource pressure, and image pull status.
- Likely path. Roll back to the last good release from the Helm chart, then fix forward.
- kubectl
- Helm
- Grafana
Certificate or DNS issue
- Signal. TLS errors, or a name that will not resolve.
- I check. Certificate expiry and issuer, cert-manager status, Route 53 records, and the ingress.
- Likely path. Renew or reissue the certificate, or correct the DNS record.
- cert-manager
- Route 53
- openssl
How a change reaches production, safely
This is the path a change takes from a developer to a validated release. Select any step to see what happens there and how I work with it. Rollback and observability are part of the flow, not an afterthought.
swipe to see the full pipeline
Stage 01 / 10
Developer
Work starts with a change. Developers open a merge request, and part of my job is helping them get there: access, environments, and answers when a build or deploy does not behave.
Case studies, not screenshots
Three pieces of work I can talk through end to end: what the problem was, what I built, and what it changed.
Homelab v2.5 - Bare-Metal Kubernetes Platform
A 22-phase, production-style Kubernetes platform built by hand on bare-metal Proxmox to understand cluster lifecycle from first principles.
- Kubernetes
- kubeadm
- Calico
- MetalLB
- Longhorn
- cert-manager
- GitLab CE
- Trivy
Zero-Downtime EKS Migration
Moved production workloads from a self-managed Kubernetes platform to Amazon EKS with no customer-facing downtime.
- Amazon EKS
- Kubernetes
- Blue-Green
- Terraform
- Helm
- AWS
GitOps CI/CD and Observability Platform
CI/CD on GitLab with ArgoCD GitOps delivery and a unified observability stack that cut mean time to recovery by 40 percent.
- GitLab CI
- ArgoCD
- Helm
- Trivy
- Prometheus
- Grafana
- Loki
- PagerDuty
The short version
The highlights a recruiter needs first. Every project on this page backs it up, and the full resume is a quick email away.
Strengths
Hands-on execution
Comfortable owning work end to end, from a Terraform module to a production cutover to the runbook that supports it.
Cost and reliability
Cut AWS spend 18 percent and reduced mean time to recovery 40 percent by making real, measured changes.
Automation first
CI/CD, GitOps, and infrastructure as code so changes are repeatable, reviewable, and auditable.
Production troubleshooting
Unified observability and on-call runbooks that shorten the path from alert to root cause.
Documentation
Runbooks and standardized charts so the next engineer can pick up the work without a handoff meeting.
Continuous learning
Building a bare-metal Kubernetes platform by hand to keep the fundamentals sharp.
Certifications
- AWS Certified Solutions Architect - Associate (SAA-C03)
- Certified Kubernetes Administrator (CKA)
Education
Bachelor of Science in Computer Science
IBAIS University, Dhaka, Bangladesh
Want a copy of my resume? Email me and I will send the latest version.
For the person deciding who to interview
Straight talk, no filler. Here is what you get, and each point maps to something concrete on this page.
I own outcomes, not tickets
I took a zero-downtime EKS migration from plan to cutover to rollback readiness on workloads where downtime is not acceptable.
I make numbers move
18 percent less AWS spend and 40 percent lower mean time to recovery are the kinds of results I look for, and I can explain exactly how each happened.
I automate the boring parts
Reusable Terraform modules, GitLab and GitHub Actions pipelines, ArgoCD GitOps, and Ansible across 50+ servers.
I know the fundamentals
AWS Solutions Architect Associate and Certified Kubernetes Administrator, plus a bare-metal cluster I built by hand to prove it to myself.
I leave a trail
On-call runbooks, standardized Helm charts, and GitOps history so the team is not dependent on one person's memory.
I work well with security
Remediated Wiz findings, wired Trivy into CI, and supported SOC 2 audit evidence collection.
Let's talk about your infrastructure
If you are hiring for a Cloud, DevOps, Platform, or SRE role, I would be glad to walk through my work. The fastest way to reach me is email.
Buffalo, New York · U.S. work authorized
Tuhin Zaman · Cloud DevOps Engineer · tuhinzaman.com