~/yogesh
Get in touch
← yogeshtiwari.com
Profile

Yogesh Tiwari

DevOps & SRE Lead in Hyderabad. About seven years across Linux, Windows Server, Kubernetes and AWS/Azure.

    Roles

    1. 01Solution architectMulti-region feasibility study for a freight platform: 13 blockers ranked, phased warm-standby design.→
    2. 02System designerThree products designed from requirements to deployment, solo.→
    3. 03DevOps engineerGated CI/CD, rollback by re-tagging an image with no rebuild, Terraform and CDK.→
    4. 04SRESLOs and 70 monitors as code; root-cause analysis of a deploy outage that blocked three environments.→
    5. 05Cloud engineerAWS end to end, plus 38 cost recommendations worth about 30% of a non-prod account forecast.→
    6. 06Platform engineerOn-prem Kubernetes at IdeyaLabs, hybrid estates, a Proxmox/KVM home lab.→

    Toolbox

    Cloud
    AWSECS FargateLambdaRDS / AuroraSQSS3CloudFrontCognitoDMSAzureFinOps
    Platforms
    Kubernetes (on-prem)OpenShift (lab)DockerProxmox / KVMLinuxWindows ServerIIS / .NET
    IaC and delivery
    TerraformAWS CDKGitHub ActionsGitLab CECodePipelinegated releasesrollback drills
    Observability and SRE
    PrometheusGrafanaLokiZabbixDatadogCloudWatchSLOs as codeRCA
    Data and backup
    PostgreSQLMariaDBRedisVeeamDuplicatiAWS BackupDR design
    Build
    PythonTypeScriptGoSQLBashFastAPINodeReactNext.jspfSense

    Timeline

    1. Jun 2026 – present

      DevOps & SRE Lead

      IdeyaLabs · Hyderabad

      Platform and reliability for Hadara, an on-premises Kubernetes platform, and delivery for a serverless AWS data pipeline.

    2. ~6.7 years, until Jun 2026

      Senior Infrastructure & Cloud Engineer

      Osmosys Software Solutions

      Linux and Windows Server across hybrid estates, backup and DR, monitoring, database operations.

    3. 2021 – 2024

      BCA, Cloud Computing

      Chandigarh University

    Certifications

    CertificationStatus
    Microsoft MS-900 · Microsoft 365 FundamentalsACHIEVED
    CKA · Certified Kubernetes AdministratorIN PROGRESS
    Microsoft AZ-500 · Azure Security EngineerIN PROGRESS
    Microsoft AZ-305 · Azure Solutions Architect ExpertEXAM SCHEDULED
    AWS SAA-C03 · Solutions Architect AssociatePLANNED

    How I work

    Eight habits that show up in every case file on this site. The right-hand column is where to check them.

    HabitWhat it means in practiceWhere it shows
    Read-only firstInventory the live account, prove the blast radius, take a backup, get approval, then change. Nothing is deleted on a hunch.Both client audits started from the live accounts, not the repositories. The tax platform's storage clean-up followed that order.
    Every change has an undoThe rollback is written and, where possible, drilled before the change runs. Releases keep the previous image tagged.Re-tag rollback in the tax platform; a production credential rotated without an outage; staging moved to ECS in small, reversible days.
    Verify, don't trust the toolA report, a diagram or a console tile is a claim. The account, the logs and the metrics are the evidence.Where code and account disagree, the case file says so, for example the CDK-defined Amplify app against the live web host.
    Say what is not builtEvery claim carries a status: delivered, in progress, planned or proposed. Dashed boxes mean exactly that.Status tables on both case files; the warm-standby region is drawn dashed because it is a proposal.
    Measure reliability, not activityAlerts exist to protect an objective. Monitors, SLOs and dashboards live in code, so they are reviewed like any other change.70 Datadog monitors and 4 SLOs defined in Terraform, alongside CloudWatch alarms and canaries.
    Name the gap to the standardFor each platform I write down where it is today, where an enterprise would hold it, and rank the gap by risk.A “Today versus the enterprise standard” table on each case file, and 13 ranked blockers to active/active.
    Cost is a reliability inputSavings are found by inventory and checked against usage, because the cheapest fix is the one that does not break anything.38 cost recommendations worth about 30% of a non-prod account forecast; snapshot storage cut by 79%.
    Write it downDecisions, runbooks and trade-offs are recorded, so the next engineer does not need me in the room.Audit trails on the products, and the numbered plans behind the client work.