Weekend Agent Challenge: Hackathon Scout Agent – Never Miss an AWS or AI Hackathon Again
Hackathon Scout Agent is an always-on AI assistant that automatically monitors Devpost for new hackathons, uses Amazon Bedrock to identify AI, AWS, and cloud-related events, and sends email alerts so you never miss an opportunity.
Introduction
Recently, I missed the opportunity to participate in the Vercel and Amazon hackathons because I wasn't aware they had opened for registration until it was too late. Like many developers and students, I don't have time to check platforms like Devpost every day, and it's easy to miss exciting opportunities when new hackathons are announced.
This experience inspired me to build Hackathon Scout Agent, an always-on AI-powered assistant that continuously monitors Devpost for newly published hackathons. Instead of manually checking for updates, the agent automatically runs on a schedule, analyzes each hackathon using Amazon Bedrock (Nova Lite), and identifies events related to topics I'm interested in, such as AWS, AI, Generative AI, Cloud Computing, and Machine Learning.
Whenever a relevant hackathon is discovered, the agent sends an instant HTML email notification to my inbox. It also stores previously processed hackathons in Amazon DynamoDB to ensure duplicate notifications are never sent. Running entirely on serverless AWS services, the agent works in the background 24/7, making sure I never miss another hackathon opportunity.
Vision & What the Agent Does
The vision behind Hackathon Scout Agent is to ensure that developers never miss valuable hackathon opportunities because they were announced without them noticing. After missing the Vercel and Amazon hackathons, I realized that manually checking platforms like Devpost every day is inefficient and easy to forget. I wanted an intelligent assistant that could monitor hackathons for me and notify me only when something relevant appeared.
The vision behind Hackathon Scout Agent is to ensure that developers never miss valuable hackathon opportunities because they were announced without them noticing. After missing the Vercel and Amazon hackathons, I realized that manually checking platforms like Devpost every day is inefficient and easy to forget. I wanted an intelligent assistant that could monitor hackathons for me and notify me only when something relevant appeared.
Hackathon Scout Agent is an always-on AI-powered assistant that runs automatically on a scheduled interval using Amazon EventBridge Scheduler. Every time it is triggered, the agent retrieves the latest hackathons from the Devpost API and uses Amazon Bedrock (Nova Lite) to determine whether each hackathon is related to topics such as AWS, Artificial Intelligence (AI), Generative AI (GenAI), Cloud Computing, or Machine Learning.
To avoid sending duplicate notifications, the agent checks Amazon DynamoDB to see whether a hackathon has already been processed. If the event is new and relevant, it generates a beautifully formatted HTML email and sends it directly to my inbox. This allows me to stay informed about hackathons without constantly visiting multiple websites.
Instead of waiting for me to search for opportunities, the agent works independently in the background 24/7. By combining scheduled automation with AI-powered filtering, Hackathon Scout Agent ensures that I receive timely, relevant notifications whenever a new hackathon matching my interests is published.
Architecture Overview
AWS Services Used
| AWS Service | Purpose |
|---|---|
| Amazon EventBridge Scheduler | Triggers the Lambda function on a scheduled interval to continuously monitor for new hackathons. |
| AWS Lambda | Executes the serverless application, scrapes the Devpost API, filters hackathons with AI, checks duplicates, and sends email notifications. |
| Amazon Bedrock (Nova Lite) | Uses AI to determine whether a hackathon is relevant based on topics such as AWS, AI, GenAI, Cloud, and Machine Learning. |
| Amazon DynamoDB | Stores previously alerted hackathons to prevent duplicate email notifications. |
| AWS Identity and Access Management (IAM) | Provides secure permissions for Lambda to access Amazon Bedrock, DynamoDB, and CloudWatch. |
| Amazon CloudWatch | Collects logs and monitors Lambda executions for debugging and operational visibility. |
How I Built It
- Defined the project architecture using Terraform to automate the deployment of AWS resources.
- Created an Amazon EventBridge Scheduler rule to trigger the agent automatically at regular intervals.
- Developed the core logic as an AWS Lambda function using Python.
- Integrated the Devpost JSON API to retrieve the latest active hackathons.
- Used Amazon Bedrock (Nova Lite) to analyze each hackathon and determine whether it matched topics such as AWS, AI, GenAI, Cloud, and Machine Learning.
- Implemented Amazon DynamoDB to store processed hackathon IDs and prevent duplicate email notifications.
- Built a responsive HTML email template to provide clear and attractive hackathon alerts.
- Configured Gmail SMTP to automatically deliver email notifications to the user's inbox.
- Applied AWS IAM roles and permissions to securely allow Lambda to access Bedrock, DynamoDB, and CloudWatch.
- Used Amazon CloudWatch to capture logs and monitor Lambda executions for debugging and operational insights.
- Created a custom
build_lambda.pyscript to automatically install Python dependencies and package the Lambda function during Terraform deployment. - Tested the complete workflow by triggering the scheduler, verifying AI filtering, checking duplicate detection, and confirming successful email delivery.
- Published the source code to GitHub for version control and reproducibility.
Challenges & Solutions
- Challenge: Avoiding duplicate hackathon notifications.
Solution: Stored processed hackathon IDs in Amazon DynamoDB and checked them before sending any email. - Challenge: Filtering only relevant hackathons instead of using simple keyword matching.
Solution: Used Amazon Bedrock (Nova Lite) to intelligently analyze hackathon descriptions and determine their relevance to AWS, AI, GenAI, Cloud, and Machine Learning. - Challenge: Packaging Python dependencies for AWS Lambda.
Solution: Created a custombuild_lambda.pyscript to automatically install dependencies and package the Lambda deployment during the Terraform workflow. - Challenge: Sending visually appealing email notifications.
Solution: Designed a responsive HTML email template and integrated Gmail SMTP to deliver formatted alerts directly to the inbox. - Challenge: Running the agent automatically without manual intervention.
Solution: Configured Amazon EventBridge Scheduler to trigger the Lambda function at regular intervals, making the agent fully autonomous. - Challenge: Monitoring and debugging the application.
Solution: Used Amazon CloudWatch Logs to track Lambda executions, monitor errors, and troubleshoot issues efficiently. - Challenge: Securely granting AWS service permissions.
Solution: Applied AWS IAM roles with least-privilege permissions, allowing Lambda to access only the required AWS services.
What I Learned
- Built an event-driven serverless application using AWS.
- Integrated Amazon Bedrock (Nova Lite) for AI-powered filtering.
- Used Amazon EventBridge Scheduler to automate tasks.
- Stored application data with Amazon DynamoDB.
- Deployed infrastructure using Terraform.
- Managed secure permissions with AWS IAM.
- Monitored application logs using Amazon CloudWatch.
- Learned how AI and automation can solve real-world problems efficiently.
Demo / Screenshots
Lambda
DynamoDB
Mail Template
Github Repository
Conclusion
Building Hackathon Scout Agent was a great opportunity to explore how AI and serverless AWS services can work together to solve a real-world problem. Instead of manually checking hackathon platforms every day, the agent continuously monitors Devpost, intelligently filters relevant events using Amazon Bedrock, and sends timely email notifications automatically.
This project demonstrates how an always-on AI agent can save time, improve productivity, and ensure important opportunities are never missed. It also strengthened my understanding of event-driven architectures, serverless development, and infrastructure automation on AWS. In the future, I plan to enhance the agent by supporting multiple hackathon platforms, personalized interest preferences, and additional notification channels.
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article