Honkerdam Octopuses: Good First Issues Recommender - promote and make contributing to OS projects easier

[anopsy]

Project Name

Good First Issues Recommender

Description

Good First Issues Recommender helps to make the first step and start contributing to open source easier.
The vast number of projects and issues can be a real blocker for people who would like to start their journey in open source.
We employ the power of collaborative filtering to make contributing easier.

Build

Yes

Train

Yes

Analyze

No

Challenge Topic / Topic Category

• Customized Information Extraction
• Education
• Enhancing Accessibility in Healthcare Through AI
• Mitigating Natural Disasters in a Changing Climate
• Reducing Food Waste Through AI-Powered Innovation
• other - World Issues
• other - Open-Source
• other - Scientific Computing
• other

Project Repository URL

https://github.com/vflores-io/honkerdam_octopuses

Deployed Endpoint URL

No response

Project Video File (not folder) Link (ensure viewer access)

https://drive.google.com/file/d/1DYKMaiIEVch8375XCLcOdmUTfEAj-zLD/view?usp=sharing

Team Members

@vflores-io , @anopsy

[naumnaum]

Your project addresses a genuinely important and relatable challenge—making the first step into open-source contributions easier. Personally, I’ve faced this very problem, and having a tool like this would have been incredibly helpful!

The current implementation is already a strong example of "Build"—it’s a tangible product that delivers value and directly solves a real-world issue.

Here are some suggestions for potential next steps:

LLM-Powered Chatbot: Introducing a chatbot to assist users with common questions on their journey to resolving their first issue could take this project to the next level. Focusing on a specific niche, such as AI inference engines or vector databases, would allow you to fine-tune the bot for maximum utility in that domain.

Advanced Similarity Models: Enhancing the issue and profile similarity analysis with more sophisticated models could significantly improve the recommendations. Leveraging vector similarity and large-scale LLMs, hosted on platforms like Modal, would provide greater accuracy and better performance. The current approach using simple fully connected layers and sigmoid activation, while functional, has clear room for enhancement.

Overall, this is an impressive and practical solution with strong potential for further growth and impact!

[anopsy]

Thank you for your valuable feedback and suggestions. It took me over three years to start contributing because I struggled to choose the right project and identify a good starting point. Our goal was to create an MVP that works, and we can now refine it further. Your suggestions align perfectly with our vision. We also plan to expand the idea and make this project the go-to resource for anyone looking to start their open-source journey.

[vflores-io]

Thank you @naumnaum for your insightful feedback. We believe this tool can significantly lower the entry barriers to open source contributions. Our goal is to make programming more accessible from the very beginning, creating a more welcoming starting point.