name: Armughan Mehboob Bhutta
role: DevOps Β· SRE Β· CloudOps Β· MLOps Engineer
location: Lahore, Pakistan π΅π°
experience: 7+ years
current_focus: [Kubernetes platform ops, GitOps with ArgoCD, DevSecOps, MLOps]
philosophy: "Automate the boring, harden the risky, make deploys boringly predictable."- π Senior DevOps Engineer β most recently at Systems LTD, one of Pakistan's largest software houses
- π§± I build golden paths: IaC everywhere, CI/CD that just works, observability you can act on
- π Deep experience with security & compliance controls β HIPAA, SOC 1/2, PCI DSS
- βοΈ Multi-cloud & on-prem: AWS, Azure, GCP, DigitalOcean, Proxmox, VMware ESXi
- π€ MLOps / AI infrastructure β deployed GPU-accelerated inference & computer-vision workloads (see below)
- π€ I mentor engineers and spread DevOps best practices across teams
| Metric | Result |
|---|---|
| π Manual release effort | ~70% reduction via GitOps CI/CD (ArgoCD) |
| π Deployment frequency | ~3Γ increase |
| β‘ Infra provisioning time | ~60% faster with IaC (Terraform) |
| π‘οΈ Service downtime | ~40% lower with service mesh + HA design |
π³ Containers & Orchestration
π‘ Observability & Messaging
π€ MLOps & AI Infrastructure
Beyond classic DevOps, I've built and operated the infrastructure that runs AI/ML workloads in production:
- π§ Built, packaged (as artifacts) and deployed a C++ AI inference engine into production
- ποΈ Deployed computer-vision applications β facial recognition, vehicle detection, and crowd-management systems
- β‘ Provisioned GPU compute with the full CUDA Β· cuDNN Β· TensorRT stack for accelerated inference
- π Ran Apache Kafka as the messaging backbone to speed up model/API communication
- π¦ Containerized GPU workloads and wired them into CI/CD for repeatable model delivery
Real-world platform engineering work, published as sanitized reference implementations.
| Project | What it demonstrates | Stack |
|---|---|---|
| π gitops-argocd-platform | App-of-Apps GitOps delivery to Kubernetes β Helm releases, sync waves, sealed secrets | ArgoCD Β· Helm Β· K8s |
| π terraform-cloud-modules | Reusable, composable IaC modules for AWS & Azure (networking, AKS/EKS, storage) | Terraform |
| π zabbix-isp-monitoring | Multi-ISP link monitoring with dynamic discovery + automated HTML email reporting | Python Β· Zabbix |
| πͺ΅ elk-centralized-logging | End-to-end log pipeline (syslog/firewall β Elastic Agent β dashboards) with TLS & auth | Elastic Stack |
| βοΈ azure-tenant-automation | One-click tenant provisioning: Function Apps, Cosmos, Key Vault, VNET via az CLI + pipelines | Azure Β· Bash |
| π€ gpu-inference-stack | Reproducible CUDA/cuDNN/TensorRT runtime for GPU-accelerated model inference | NVIDIA Β· Docker |
| βΈοΈ helm-charts-for-k8s | Reusable Helm charts for application delivery | Helm Β· Kubernetes |
Β
2025 Β· 37E2A15ED3559E26Β
2022 Β· 854SGJ0DF2R1QVSTΒ
2020 Β· 200-145-001Β
2020 Β· 17834571Β
2018ΒΒ
2017
β From armghan β let's build reliable platforms.


