Skip to content
Back to journey

// experience

Software Engineering Intern

Software engineering internship at a startup, working across backend authorization systems, reliability, and DevOps: policy-based access control, production AWS and Kubernetes infrastructure, Terraform-based IaC, observability, and alerting.

Apr 2026 – Present
AWSEC2IAMCloudWatchSNSLambdaTerraformK3sKubernetesPrometheusGrafanaOPAFluxCDGitHub Actions

Backend & Authorization

  • RBAC Policy Engine

    Designed a role-and-scope authorization engine in Open Policy Agent — 9 roles across 5 scope levels, with deny rules enforcing separation-of-duties.

    Write-up

  • IAM-Style Policy Attachment

    Built an AWS-IAM-style engine in OPA: per-user inline and attachable role policies with Allow/Deny precedence and live-binding resolution.

    Write-up

  • Notification System

    Built multi-channel notification delivery and a compliance activity feed derived directly from notification events.

  • Workspace & Collaboration

    Built workspace chat with mentions and notifications, plus activity tracking for uploaded evidence.

Infrastructure & IaC

  • AWS Infrastructure

    Hands-on with EC2, IAM, CloudWatch, SNS, and Lambda in production.

  • Terraform & IaC

    Wrote and reviewed Terraform plans; caught one that would have restarted every production node.

  • K3s & Kubernetes Ops

    Managed production K3s nodes — scheduling, taints, draining, resource allocation.

Reliability & Observability

  • Kubernetes Reliability

    Closed a CoreDNS availability gap with HPA/PDB enforcement and topology spread constraints.

  • Monitoring & Observability

    Designed Prometheus monitoring around the four golden signals, using real traffic data to pick probe targets.

  • SLO Monitoring

    Built SLI/SLO/SLA-based health checks and the alerting rules on top of them.

Incident Response

  • Alerting Pipeline

    Verified a CloudWatch → SNS → Lambda → Slack alert pipeline end-to-end in production.

  • Production Troubleshooting

    Investigated live incidents — config mismatches, deployment issues, resource problems.

  • Database Reliability

    Root-caused a MongoDB liveness-probe issue causing restarts under node load.

Delivery & Architecture

  • Deployment Integrity

    Discovered production was running an unmerged branch and re-cut it against live config.

  • Cost-Aware Architecture

    Evaluated cost-effective monitoring architectures and flagged a CloudWatch billing trap.

  • CI/CD

    Built self-hosted GitHub Actions runners with cross-architecture (x64 → arm64) builds.

    Write-up