Prashant Kumar Maurya
Site Reliability Engineer
Site Reliability Engineer focused on production-grade Kubernetes platforms: multi-environment GitOps delivery, progressive rollouts gated by SLOs, self-healing automation, and infrastructure as code. I like building systems that recover themselves and leaving cloud accounts spotless.
Experience & Projects
Atlas Platform — Multi-Environment GitOps on AWS EKS
GitHub- Three namespaces (three-tier-dev, three-tier-staging, three-tier-prod) on one EKS 1.31 cluster in ap-south-1, reconciled by Argo CD ApplicationSets with sync waves -1 through 3.
- Argo Rollouts canary on the backend Rollout: 25% traffic, Prometheus AnalysisTemplate, promotion steps at 50/75/100; abort when success rate drops below 0.99 for two consecutive measurements.
- CloudNativePG Cluster (3 instances), PgBouncer, daily backups and WAL archive to per-env MinIO; PITR and failover exercises documented in runbooks and INC postmortems.
- Terraform modules for VPC, EKS, and IAM/IRSA; destroy.sh runs top-down Kubernetes cleanup before terraform destroy to avoid orphaned EBS volumes.
Tools: Kubernetes (EKS), Terraform, ArgoCD, Argo Rollouts, Kustomize, CloudNativePG, Helm, Prometheus, Loki, Grafana, NGINX Ingress, k6, AWS, Bash
GitOps Drift Shield — Auto-Healing SRE Pipeline
GitHub- Python drift-detector runs terraform plan -json -detailed-exitcode; exit code 2 means pending changes.
- GitHub Actions drift-check.yml cron every 30 minutes; Jenkins applies saved plan and writes audit-logs/*.json to the audit bucket.
- Ansible site.yml configures Flask demo and node_exporter on EC2; optional K8s monitoring namespace runs Prometheus, Pushgateway, Grafana.
Tools: Terraform, Ansible, Python, AWS (EC2, S3, VPC), Kubernetes, Jenkins, Prometheus, Grafana, GitHub Actions, Slack, Bash
Sentinel — Auto-Healing Infrastructure & CI/CD
GitHub- sentinel-monitor polls APP_URL every 5s and restarts CONTAINER_NAME via /var/run/docker.sock when /health fails.
- deploy.yml provisions EC2 in ap-south-1, reads server_public_ip, runs ansible/setup.yml with dynamic inventory.
- Terraform state lives in S3 with DynamoDB locking; CI sets terraform_wrapper: false so Ansible receives an unmasked EC2 IP.
Tools: Python, Docker, Docker Compose, Terraform, Ansible, AWS (EC2, S3), GitHub Actions, Bash
PromptTune — AI Prompt Engineering Assistant
GitHub- src/app/actions.ts streams Gemini responses with per-user rate limits and attachment URL allowlists.
- Client uses Firestore onSnapshot for chats; Firebase Storage for images and PDFs with client-side compression.
- Deployed on Vercel with GEMINI_API_KEY and Firebase public config in environment variables.
Tools: Next.js, TypeScript, Gemini API, Firebase, Tailwind CSS
Technical Skills
Cloud & Infrastructure: AWS (EKS, EC2, S3, VPC), IAM / IRSA, EBS, NAT Gateway, Security Groups, Terraform modules, Remote state, S3 state-locking
Kubernetes & GitOps: Kubernetes (EKS, kind), CRDs & Operators, RBAC, Secrets, Sync Waves, ArgoCD App-of-Apps, ArgoCD ApplicationSets, Argo Rollouts, Kustomize, Helm, Docker, Docker Compose
CI/CD & Automation: GitHub Actions, Jenkins Pipelines, Groovy, Ansible playbooks, Dynamic inventory, Bash scripting
Observability & SRE: Prometheus recording rules, ServiceMonitors, Pushgateway, Grafana, Loki, Alertmanager, SLO / SLI design, Multi-window burn-rate alerts, k6 load testing
Databases & Languages: PostgreSQL, CloudNativePG, HA & PITR, WAL archiving, PgBouncer, MinIO, Python, Go, YAML / JSON, Git, AWS CLI, kubectl, jq
Community & Open Source
Agentic AI Foundation (AAIF), Linux Foundation
Community Event Organizer — Delhi NCR
Running and coordinating local community events for the Delhi NCR region, connecting engineers around cloud-native and agentic AI topics.