Project Never Done

Inspiration

The Tutorial and other big data applications.

What it does

Offer a complete website where you can upload log files that will be analyzed and can be asked questions.

How we built it

Build a independent frontend and backend and then connected it. The backend uses the pretrained model "all-MiniLM-L6-v2" for embedding and "gpt-3.5-turbo" for answering.

Challenges we ran into

Handling a large log file and designing a responsive frontend.

Accomplishments that we're proud of

Creating a working system that gives satisfactory results.

What we learned

How to use NLP methods to handle large data and how to design a system around it.

What's next for Project Never Done

Further development on the system would mean doing a more concrete log analysis with domain knowledge to reduce the token amount and training a custom LLM for answering more specific domain questions. GitLab link: git@gitlab.com:rs_hackathon/hackathon34.git

Built With

• llm
• openai
• python
• react