Architecture Evolution of a WordPress Site
A single EC2 box running WordPress and MySQL, evolved step by step into a highly available, auto-scaled deployment with RDS and EFS — a diagram for every stage.
Mechanical engineer turned cloud builder — a year of deep AWS study and a stack of things I've actually deployed.
About
I have a B.S. and M.S. in Mechanical Engineering from Stony Brook University (GPA 3.94), then spent several years leading a multi-site retail business in Georgia — hiring and training staff, running full-scope renovations, and growing one location to six-figure monthly sales before a $2.4M sale. A different world from cloud, but both taught me to reason about systems under real constraints.
About a year ago I committed to a switch into cloud engineering. Since then I've earned four AWS certifications, joined the year-long Amazon Cloud Institute developer program and the BeSA Bedrock AgentCore cohort, and built and deployed the projects below — this site included, hosted on AWS Amplify with Route 53 DNS.
Projects · deployed
All projects →A single EC2 box running WordPress and MySQL, evolved step by step into a highly available, auto-scaled deployment with RDS and EFS — a diagram for every stage.
Rehosted a running Linux web server into AWS with Application Migration Service — continuous block-level replication, a validated test launch, and cutover to production in under 25 minutes, reconnected to its database on RDS.
One Python Lambda handles create, read, update, delete and list against DynamoDB, exposed over HTTPS through API Gateway. Ships with a CloudFormation template that stands up the whole stack in one deploy.
n8n running in Docker on a free-tier EC2 instance, served over HTTPS at n8n.builtbydhruv.com — Caddy handles TLS via Let's Encrypt, Route 53 DNS, and a nightly S3 backup script with a documented 24-hour RPO and 15-minute RTO.
A static HTML/CSS résumé hosted on S3, served over HTTPS through a CloudFront distribution for TLS and CDN performance. CloudFormation template included.
Building & learning now
Capstone: an event-driven identity-verification backend for a bank. Rekognition compares a selfie against a driver's license photo at an 80% similarity threshold, Textract cross-checks 8 license fields against the application, and DynamoDB auto-scales 2–20 capacity units at 70% utilization. Next: splitting it into async Lambdas orchestrated with Step Functions.
Building AI agents with n8n and Google Gemini to replace Wayfair vendor management's manual trend and competitor research, delivering a live dashboard of insights to support a go/no-go decision on automating the workflow.
Hands-on Amazon Bedrock AgentCore: building agentic-AI proof-of-concepts and taking them toward production.
AWS Cloud Application Developer — year-long AWS-run program covering architecture, compute, networking, security, Python and generative AI, with a capstone.
Andrew Ng / DeepLearning.AI & Stanford (Coursera) — supervised learning, neural networks, and the math underneath.
Skills
Contact
The fastest way to reach me is pdhruv331@gmail.com — or connect on LinkedIn.