
Weekend Creative Agent Challenge: AWS Daily Topic Agent — CLF-C02 para el SBG Univalle
"I built an autonomous agent that generates a new AWS CLF-C02 topic every day for my university community — no manual intervention, just EventBridge + Lambda + Bedrock doing their thing. The best study tool is the one that's already ready when you open it. 🚀"
As a systems engineering student and leader of the AWS Student Builder Group at Universidad del Valle Sede Tuluá, one of our biggest challenges is keeping our community consistently learning about AWS services. Students preparing for the AWS Cloud Practitioner (CLF-C02) certification often don't know where to start or what to study each day.
That's why I built the AWS Daily Topic Agent — an always-on autonomous agent that automatically generates a new AWS service topic every day for our community. Every morning, without anyone opening an app or running any script, the agent wakes up, asks Amazon Bedrock to generate an educational topic focused on the CLF-C02 exam, and publishes it instantly to our community web page.
The page shows the service of the day, a clear description, a real practical example, and a key exam tip — everything a student needs to learn something new in just a few minutes.
Vision and what it does
The AWS Daily Topic Agent solves a real problem for our community: consistency. It's hard to study every day when you have to decide what to learn. This agent removes that friction entirely — every day there's already a new topic waiting, generated automatically by Amazon Bedrock and published to our web page without any human intervention.
The agent picks a different AWS service each day from the CLF-C02 domain, explains what it is, gives a real-world practical example, and includes a specific tip for the certification exam. Our community members just open the page and start learning.
How I built it
I started by creating an S3 bucket configured as a static website to host both the web page and the daily JSON file generated by the agent.
Then I created a Lambda function in Python that connects to Amazon Bedrock using the Nova Lite model. The function sends a structured prompt asking Bedrock to generate a JSON with four fields: the AWS service name, a simple description, a practical use case, and a tip for the CLF-C02 exam. Once generated, the Lambda saves the JSON directly to S3.
The biggest challenge was getting the correct request format for the Nova Lite model through the Bedrock API. The content field requires a specific array structure that took several iterations to get right. At first I was sending the prompt as a plain string, then as an object with a type key — both were rejected. The correct format turned out to be an array with a single object containing only a text key.
For the frontend I used Kiro to generate a clean, modern HTML page with AWS colors — dark theme, orange #ff9900 — that fetches the JSON from S3 on every visit and renders the daily topic dynamically. I also had to configure CORS on the S3 bucket so the page could read the JSON file from the browser correctly.
Finally I connected everything with EventBridge Scheduler using a cron expression (0 13 * * ? *) that triggers the Lambda every day at 1:00 PM Colombia time, so the community always has a fresh topic ready when they check the page.
AWS services used and architecture
- Amazon S3 — static website hosting and JSON storage
- AWS Lambda (Python 3.12) — orchestration and Bedrock invocation
- Amazon Bedrock (Nova Lite) — daily educational content generation
- Amazon EventBridge Scheduler — daily automatic cron trigger
- AWS IAM — permissions for Lambda to access Bedrock and S3
Architecture flow:
EventBridge Scheduler → Lambda → Amazon Bedrock → S3 → Web Page
The entire stack runs on AWS Free Tier eligible services, with an estimated cost under $2/month.
What I learned
This challenge taught me how to connect multiple AWS services into a fully serverless, event-driven workflow with zero manual intervention. Specific things I learned:
- The exact request body format required by Amazon Nova Lite through the Bedrock invoke_model API
- How to configure CORS on an S3 bucket to allow browser fetch requests from a static website
- How to use EventBridge Scheduler with cron expressions to automate Lambda invocations
- How IAM roles and policies work in practice — specifically attaching AmazonBedrockFullAccess and AmazonS3FullAccess to a Lambda execution role
- How to store and serve dynamic JSON content from S3 as a lightweight backend
Most importantly, I learned that you can build something genuinely useful for a real community — not just a demo — using only Free Tier AWS services in a single weekend.
Live app
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article