Inspiration
Increasingly organizations are mandated by regulations to provide their ESG data as part of their yearly reporting. With regulations, organizations are also at risk of paying penalties for any violations, for e.g., upto 5% of global revenues in EU. Today, what organizations lack is the feedback loop from operations regarding their ESG mandates to avoid penalties.
What it does
Sy bridges the knowledge gap between different operations in an organization, how they affect each other and how different regulations are required to be met against these operating activities.
How we built it
Input corpus data: Regulations from Public Domain, Project activities and metrices, Mock ERP data Tools: Groq, Llama, LlamaIndex We hosted LLM as a service from Groq. We built a RAG using the corpus data and generated vector search index backed by chromadb. Then we hosted this in Django backend server to query upon. We also attempted to build a knowledge graph index
Challenges we ran into
Limited operations dataset
Accomplishments that we're proud of
This was first exposure to LLM models. We are glad that we were able to go through multiple iterations and test our idea building our knowledge around LLM through networking.
What we learned
Concepts of LLM, Deploying LLM, Dataset requirements, Limitations of LLM
What's next for Sy
Refine the product architecture to build the MVP
Built With
- groq
- llama
Log in or sign up for Devpost to join the conversation.