AWS Cloud & DevOps Engineer · Career Changer · South Fulton, TN · Open to Remote & Memphis/Paducah Area Roles

I spent 15 years running automated control room systems in manufacturing solo overnight operations, no supervision, meticulous documentation, zero tolerance for downtime. Now I'm applying that same mindset to cloud infrastructure and AI agent systems. One year into an intensive AWS/DevOps bootcamp (completed), 26+ public projects, and currently beginning a Bachelor of Science in Artificial Intelligence.

Cloud AWS Lambda · S3 · EC2 · RDS · VPC · IAM · Route 53 · CloudFront · API Gateway · SNS · SQS · Auto Scaling · Aurora Serverless · Secrets Manager · Kinesis · Firehose · Glue · Athena · SageMaker · CloudWatch · X-Ray · CloudTrail

AI / ML & Agent Systems Anthropic SDK · Claude Code · FastMCP / MCP Protocol · Amazon Bedrock · SageMaker · Amazon Rekognition · Hermes (Nous Research) · Ollama (local LLMs) · RAG Pipelines · Tool Calling · Agent Loop Architecture

Infrastructure as Code Terraform · AWS CloudFormation · AWS SAM · Ansible

Containers & Orchestration Docker · Kubernetes · K3s · Helm · k9s · FastAPI

Languages Python (Boto3 · FastAPI · Typer · Strawberry GraphQL) · Bash · HCL · Go (learning)

DevOps & Observability GitHub Actions · Git · AWS CLI · Linux (Ubuntu / WSL2) · Nginx · Gunicorn · tmux

AI DevOps Triage Agent An autonomous AI agent that ingests CloudWatch alerts, reasons over them using tool-calling, and takes action Lambda invocations, ECS restarts, log queries without human intervention. Built on the Anthropic SDK with FastAPI, Lambda, Terraform, and MCP. This is the kind of agent you'd actually run in production. Anthropic SDK FastAPI Lambda Terraform MCP CloudWatch Python

FastMCP Server for AWS Observability A Model Context Protocol server that exposes AWS observability tooling (CloudWatch, X-Ray, CloudTrail) as callable tools for AI agents. Connected to Claude Desktop. Built with FastMCP. FastMCP MCP CloudWatch X-Ray Python

Minimal Agent Loop Implementation A clean, minimal implementation of an AI agent loop with real tool use built to understand the mechanics of agentic systems from first principles before scaling to larger projects. Anthropic SDK Tool Calling Python

End-to-End MLOps Pipeline on AWS Complete Full ML lifecycle on AWS: automatic training, evaluation with a performance gate, model deployment, and production monitoring entirely defined in Infrastructure as Code. SageMaker Python Jupyter IaC Model Monitor

Bedrock AI Model Comparison API Fully serverless GraphQL API that compares AI model responses side-by-side using Amazon Bedrock, Lambda, and DynamoDB. GraphQL Strawberry Lambda DynamoDB Amazon Bedrock

Multi-Region Active-Active Infrastructure Team Collaboration Vue.js application with AWS API Gateway built with bootcamp classmates. My personal contribution: Aurora Serverless with Secrets Manager credentials across two AWS regions with automated failover. Aurora Serverless Secrets Manager Multi-Region API Gateway

RAG Chatbot with Amazon Titan & FAISS Retrieval-Augmented Generation pipeline using Amazon Titan Embeddings, FAISS vector store, and S3 for document storage. Titan Embeddings FAISS S3 Python

Personal Knowledge CLI A Typer-based command-line knowledge base backed by SQLite. Built for capturing technical notes, commands, and references fast from the terminal, where I live. Python Typer SQLite

Event-Driven Image Processing Pipeline Decoupled microservices pipeline using S3 triggers, SNS fan-out, SQS message queuing, and Lambda processing a real-world async workload pattern. S3 SNS SQS Lambda Python

Terraform 3-Tier Application Production-style 3-tier AWS architecture (web, app, data) provisioned entirely with Terraform VPC, subnets, ALB, EC2 Auto Scaling groups, security groups. Zero manual console clicks. Terraform EC2 VPC ALB Auto Scaling HCL

Ansible Automation Stack Automated Flask web application deployment using Ansible playbooks with Gunicorn as the WSGI server and Nginx as the reverse proxy. Idempotent, repeatable, production-style provisioning. Ansible Flask Gunicorn Nginx Jinja

Job Tracker App on Kubernetes FastAPI-based job application tracking service containerized with Docker and deployed on Kubernetes pod management, service exposure, full deployment config. FastAPI Docker Kubernetes Python

• crossplane-contrib/crossview - PR #219 merged: modified the release workflow to replace automatic push triggers with a controlled workflow_dispatch input
• EnvSync K3s - PR #16 submitted

Former: Control room operator running fully automated feed mill systems
        solo overnight 15 years, zero supervision, meticulous documentation

Now:    Building cloud infrastructure and AI agent systems with the same
        discipline and attention to uptime that kept production running
        every night while working 40 hours a week and raising two kids

Based:  South Fulton, TN (Tennessee/Kentucky border)

Target: Remote cloud/DevOps roles or Memphis/Paducah area in-person

Email:  jmac052002@gmail.com

"Same discipline. Different stack."