🤖 MLOps Engineer • 🚀 Forward Deployed Engineer • ☁️ Cloud Architect
LinkedIn • GitHub

📰 The Bug that No AI Agents Caught 📰 How I Deployed Terraform to Azure with Zero Secrets

A TypeScript payment service built to stress-test AI code review agents. Benchmarked Greptile, CodeRabbit, Qodo Merge, and GitHub Copilot against planted business logic bugs — including a timezone-dependent settlement flaw none of them caught.

🤖 AI Benchmark | 💳 Payment Infra | 🧪 Bug Detection | 🔍 Multi-Agent Comparison

Terraform modules with CI/CD using GitHub Actions + OIDC. Key Vault integration for secrets & model artifacts. Federated identity. Built for reproducible ML pipeline deployments.

🧠 ML-Ready Infra | 🔐 Zero-Secrets Pipeline | 🏗️ Modular IaC | ☁️ Enterprise Patterns

A React-based frontend app with API search and filtering. Deployed via Netlify.

🧪 Responsive UI | 🔍 TMDb Integration | ⚙️ React + Hooks | 🚀 Full-Stack Demo

• 📜 From DevOps to MLOps: Bridging the Model-to-Production Gap
• 📜 Observability for ML Pipelines: Prometheus + Grafana in Practice
• 📜 Forward Deployed Engineering: Shipping ML at the Customer Edge

# Ship models, not notebooks.
# Observability is the bridge between training and production.
# Infrastructure should be as reproducible as your experiments.
# The best ML system is the one your customer can actually use.

Sometimes even LLMs get stumped. This case study explores how human curiosity and context outperformed AI in debugging complex Terraform issues. Full story on LinkedIn

Read the full case study on Medium

I'm open to:

• 🤖 MLOps pipeline design & model serving infra
• ☁️ Cloud architecture for ML workloads
• 🔍 AI agent evaluation & benchmarking
• 📦 OSS Terraform modules for ML platforms
• 🧑‍🏫 Mentorship or speaking invites

📫 Email • 🧠 GitHub • 🌐 LinkedIn

"The best ML infrastructure is invisible — your model ships, your pipeline scales, and your customer never knows the complexity behind it." ⚙️