Protogen

Inspiration

       Our inpiration for Protogen, for the GenAI4HUMANS stems from when we were discussing ideas. One of our teammates, Rishi, proposed the idea for making an AI to help with researchers. He'd done a research project on Alzheimers, and in the process he learned about a protein called Tau, which is a leading cause for Alzheimers. He stumbled across PyMOL, which makes models for proteins, one being Tau. He told us what he didn't like about it, but we were already invested in his story. So we got to creating the perfect tool to do just that.

What it does

        Protogen uses an AI chatbot to take user prompts and make terminal commands to put into PyMOL based on the prompts.

How we built it

        We built it using Windsurf(website generator on VS code) and an OpenAI agent. We used the backend language Python and the framework as Flask. 

Challenges we ran into

        Implementing the agent and API key would keep leading to terminal errors, one after another, but eventually once we were able to understand the terminal errors, we would be able to solve them. We had a rough time finding LLMs on Huggingface, but eventually the agent along with Windsurf's ability to create AI helped us with the LLM. We also had many problems with connecting our Windsurf Projects to our GitHub repo, but through understanding the terminal, we were able to make it work. 

Accomplishments that we're proud of,

        We're proud of us being able to pull through, even with terminal being so uncooperative. We were able to flesh out this idea that we cared about, and make it exactly what we wanted it to be. 

What we learned

         We learned how agents work, and how Windsurf operates when it comes to AIs. Just working on one feature seems to be the best thing to in a hackathon, instead of multiple features like we did in previous hackathons.

What's next for Protogen

Built With

• flask
• python
• windsurf