Netwurk

Front page
Contact Book

Inspiration

Trying to expand our networks. Hearing other people doing this manually, and wanting

What it does

You write a list of filters, click submit, and the app will research for people that match the criteria. You can click on each one, where you can see information about them, such as their linkedin, their GitHub, research areas and papers, personal website and more. You can save the person to your contact book, and from there you are able to network with them, research them for questions, and expand your horizons.

How we built it

We split up our tasks by UI, AI and overall structure. We all contributed to all parts, but had a main focus. UI was figuring out how the application was going to look and the user flow. AI was testing and reimplementing the search algorithm. The overall structure, included setting up the supabase and how each component would connect with each other.

Challenges we ran into

When we first ran our searching algorithm, based on purely what AI wrote it, it ate up 5 dollars worth of api credits in just a few test cases and only outputted 2-3 people per test case (incorrectly as well). Then we spent a healthy amount of hours, reimplementing the algorithm (by using cheaper models, including another api key without the ai layer for searching, and optimizing the prompts). At the end it was costing ~0.2 cents and outputting ~10 results. Discussing the user flow was also important, we talked about less friction and making the application easy to use. We had disagreements, but we managed to figure out something we all agreed with.

Accomplishments that we're proud of

Search algorithm efficiency. User flow. Putting all our components together.

What we learned

How to work with AI apis, and test them continuously. Small changes, and a lot of tests really allowed us to have something that worked better.

What's next for Netwurk

Netwurk allows for connection, BUT, as Randy said in the opening ceremony, DATA is extremely important. This app gives data on a group of people (job seekers, people looking to meet others). This data can find patterns and highlight problems, which can be very useful in a variety of situations. So marketing this would be the next step, by keeping it free and displaying it in events and social media.