Hackd

Front Page of HACKD
Sentiment Analysis for HACKERS
HACKER Board plus possible teammates

Inspiration

Finding teammates can be incredibly stressful. The Hack the North Slack had 1,710 members in the "looking for teammates" channel, and taking the step to message someone when you're a solo hacker can be terrifying. Everyone on our team had to take that leap, but some can't.

That's why we created Hackd—an easier way to find the right hacker for the job.

What it does

Hackd analyzes your GitHub profile to understand what kind of coder you are and the complexity of the projects you've worked on. It also conducts sentiment analysis to identify people who can help meet your goals.

But we didn't want to stop there. Recruiters come to hackathons to meet thousands of people. With Hackd, we can help them find the best hackers to join their teams or contribute to exciting projects.

How we built it

BACKEND

Convex to run our database and store the user information, this includes authentication for github and all the information we gather from their account, the sentiment analysis linked to the github account, as well as the AI algorithm results of matches
Cohere Would run the analysis of the github results and the sentiment scores to see who the matches were
Python to scrape github users of their public repos, tech stack, and profile pictures FRONTEND
React for our front-end development ...

Challenges we ran into

We had to learn new technology very fast here, Convex's dashboard and AUTH system gave us a lot of issues on figuring out the linking.

Being able to manipulate Cohere to be able to read through such heavy files such as a git repo was also quite a challenge, we had to make an insanely fine-tuned model to ensure that it avoided all unnecessary information to speed up processing.

Accomplishments that we're proud of

We completed the project !! Hackd is polished and well made web app that is able to connect users attending hackathons with each other and ensure a more clean way to find your next collaborator.

Being able to offer a solution that is DEI friendly, our app is able to connect users without the walls, until you find the person for the job you don't even know their name, and the lovely part about github is that once we show their icon it automatically isn't them. All they know is their work, work ethic, and how much they want to go.

What we learned

I think we should start with collaboration, this Hackathon was a physically demanding challenge and we pulled through together as a team, we all have each other to thank for it and brought us all closer.

Looking solely at the techstack, we learned a lot about Convex and different data base companies out there, how they work and how we can manipulate them into our product.

Parsing github was a huge learning curve, as well as how easy it is to get information out of it, the company wants us to have it, it's up to us to make it as beneficial for others as possible.

What's next for Hackd

Continuing the recruiting side of it. I believe we have a lot to grow in people able to connect top talent at hackathons with users, this includes being able to invite for a coffee-chat and having access to information like a resume at a click of a button.

As this is a proof-of-concept our AI is not perfect, even with our fine-tuning it is still slow especially in larger repos. Finding a more efficient way to go through and grade it.

We have also noticed at hackathons that there is no way to give feedback to teammates after. We think that kind of knowledge is a powerful thing that we should be able to do. A post hackathon, or coworking review could allow our algorithms to sort even better