AWS Builder Center

Weekend Creative Agent Challenge: WeatherVerse

A structured interaction between Bedrock and Lambda functions and send mail resumes. A new approach for me.

Work as a data engineer

Vision & What It Does

In our fast-paced daily routines, we often rely heavily on proactive searches—manually opening applications to check the weather, news briefings, or daily updates. However, the true pinnacle of modern automation is the invisible tool: the one you never have to open because it works seamlessly in the background. To explore this concept, I built WeatherVerse, an autonomous, always-on serverless agent designed to blend real-time meteorological data with generative AI to craft a unique daily poem delivered directly to your inbox.
The problem this agent solves is the friction of passive consumption. Instead of hunting for daily inspiration or routine updates, WeatherVerse acts as a quiet companion that evaluates environmental variables and transforms raw data into creative expression autonomously. While this specific iteration creates weather-themed poetry, the foundational architecture scales seamlessly to broader use cases. For instance, this same event-driven pattern can be adapted to summarize breaking world news, curate personalized tech newsletter digests, aggregate financial market trends, track GitHub repository activity updates, synthesize social media sentiment, or automatically draft daily project status reports.

How You Built It

Developing WeatherVerse was an exercise in keeping architecture lightweight, robust, and cost-efficient. The journey began by establishing an automated trigger mechanism using Amazon EventBridge to execute a serverless function on a regular cadence. Inside the compute layer, a Python script executes an HTTP GET request to the public Open-Meteo API to capture the exact current temperature and atmospheric indicators for a targeted coordinate.
Console AWS with Lambda code

Once the meteorological context is retrieved, the script dynamically injects that data into a custom prompt and invokes Amazon Bedrock, leveraging the cutting-edge Amazon Nova Micro foundation model. The model interprets the prompt and generates a fresh, creative poem. Finally, the system leverages Amazon SES (Simple Email Service) to format and dispatch the resulting prose straight to a verified inbox.
Mail in portuguese sent to my mailbox

The primary challenge encountered during development was managing tight execution constraints and strict payload structure rules required by foundation model APIs, which was successfully resolved by properly formatting request parameters and scaling the Lambda function's timeout configuration.

AWS Services Used / Architecture Overview

WeatherVerse relies entirely on a lean, serverless, and cloud-native architecture hosted on AWS:
  • AWS Lambda (Python 3.11+): Serves as the core compute runtime executing the end-to-end automation logic.
  • Amazon Bedrock (Amazon Nova Micro): Powers the intelligent text generation layer, interpreting prompts and outputting creative prose.
  • Amazon SES (Simple Email Service): Acts as the secure delivery mechanism, pushing the generated output to the user's email.
  • Amazon EventBridge: Handles the event-driven or scheduled triggering mechanism to ensure the agent runs completely autonomously.

Architecture Flow

Process flow designed using draw.io


What I Learned

This challenge marked my very first experience utilizing Amazon Bedrock, and what stood out most was the incredible ease of use. Coming into the cloud ecosystem, I initially anticipated a steep learning curve when attempting to integrate advanced generative AI capabilities into backend workflows. Instead, I discovered that invoking state-of-the-art foundation models like Amazon Nova through Python and Boto3 is remarkably straightforward and intuitive, requiring only a few lines of clean code.
Furthermore, the cost efficiency was truly astonishing. Even without an active Free Tier credit allocation, deploying, testing, and running this solution incurred virtually zero noticeable cost due to AWS's pay-as-you-go pricing model and the generous permanent free tiers provided by AWS Lambda and Amazon SES.
My biggest overarching takeaway is just how simple, fast, and accessible it is to combine serverless functions with foundation models. Within an extremely short timeframe, I was able to transition from an abstract architectural concept to a fully functional, autonomous production-ready agent delivering tangible, high-value end results without friction.

Link to App or Repo

You can explore the complete source code, deployment files, and setup documentation in the public GitHub repository:
Any opinions in this article are those of the individual author and may not reflect the opinions of AWS.
Enjoyed reading this content? Let the author know!

Your likes, comments, shares, and saves help creators reach more builders.

Loading recommendations

Loading article