← All Posts

From Claude to Cloud: How I Automated My Website Deployment

2026-03-01

🚀 How I Built My Website in a Day Using AI: A Mindset Shift

I have been getting a lot of questions about how I built my website in a single day using AI. As much as it surprises many, I am no different! I never truly imagined how AI could make you this productive and make your work feel so valued until I tried a hands-on project.

The secret isn't just technical; it’s about being empathetic, creative, and thinking through a solution before touching the keyboard (much like how we used to "dry run" code manually before running it on a machine).

👵 Explaining AI to My Grandma

If I had to explain this process to my grandma, I’d tell her it’s like having a very fast, very eager apprentice in the kitchen.

  • Brainstorming > Ordering: You don't just "order" a website like a pizza. You have a conversation. You explain the vibe, the goal, and who is coming over for dinner.
  • The "Dry Run": Before we even turned on the "stove" (the computer), I had to do a manual dry run in my head. If you can't explain the logic of your recipe to yourself, the apprentice is going to burn the toast.
  • Check the Seasoning: Instead of blindly accepting what the AI does, you have to taste the sauce. Be mindful, check the code, and understand why it’s doing what it’s doing. It’s a brainstorming session, not a magic trick.

🛠 The Tech Stack (The Simple Ingredients)

Based on my repository, here are the tools I used to get the job done:

  • Python & Flask: The "brain" of the operation that handles the logic.
  • HTML/CSS: The "skin and clothes" that make the site look professional.
  • Docker: Think of this as a perfectly packed suitcase. It ensures the website works exactly the same on my laptop as it does on the Google servers.

🏗 The Architecture: From Laptop to Live

I used Visual Studio Code as my workbench. The flow is simpler than you’d think:

  1. The Push: I write code and "push" it to the Git master branch.
  2. The Factory: This triggers Google Cloud Build, which automatically starts "baking" the new version of the site.
  3. The Home: The site is hosted on Cloud Run. It’s brilliant because it only "wakes up" when someone visits the site, making it incredibly efficient.

🧠 The "Context Window" Secret

I used Claude Code to accelerate the build. The real "unlock" here is the Context Window.

Imagine if your apprentice forgot the first half of your sentence by the time you reached the end. That’s a small context window. Claude, however, has a massive "short-term memory." It can "read" my entire project at once—remembering my branding, my previous errors, and my goals—so I never had to repeat myself.

💻 Technical Deep Dive: The Blueprint

For my fellow engineers, here is the cloudbuild.yaml snippet that automates my life. It moves code from my laptop to the world in seconds:

steps:
# Step 1: Build the container image
- name: 'gcr.io/cloud-builders/docker'
  args: ['build', '-t', 'gcr.io/$PROJECT_ID/anwesha-me', '.']

# Step 2: Push the image to the Registry
- name: 'gcr.io/cloud-builders/docker'
  args: ['push', 'gcr.io/$PROJECT_ID/anwesha-me']

# Step 3: Deploy to Cloud Run
- name: 'gcr.io/[google.com/cloudsdktool/cloud-sdk](https://google.com/cloudsdktool/cloud-sdk)'
  entrypoint: gcloud
  args:
  - 'run'
  - 'deploy'
  - 'anwesha-me'
  - '--image'
  - 'gcr.io/$PROJECT_ID/anwesha-me'
  - '--region'
  - 'australia-southeast1'

🧗 Challenges & Learning Curves

Even with a super-apprentice, it wasn't all smooth sailing. Here were my biggest hurdles:

The "Yes Man" Trap: AI is incredibly eager to please. If you ask for a complex solution, it will give you one, even if a simple if statement was enough. I had to learn to say, "Let’s keep it simple."

Requirement Drift: If my initial "brain dump" wasn't clear, the AI would hallucinate features I didn't need. I realized that clear requirement gathering is 80% of the battle.

Resisting Blind Trust: It's tempting to just copy-paste and pray. The real challenge was staying mindful and reviewing every line to ensure I actually understood the logic being implemented.

💡 The Big Takeaway

My biggest realization? AI doesn't replace the engineer; it elevates you. You stop being the person who just "lays the bricks" and you become the Architect and Director.

When you combine technical "dry runs" with a bit of creative empathy for your user, you aren't just writing code, you are building a solution. You don't need a decade of experience to start; you just need to be clear about what you want to create.

Don't just build code, build a vision that makes you feel valued.