MSME LOAN ADVISOR

Prerequisites

1. create account in OpenAI Platform - https://platform.openai.com/docs/overview
2. create account in Langsmith - https://www.langchain.com/langsmith

How to setup locally?

1. Download Ollama - https://ollama.com/download
2. Run in terminal:

       ollama pull llama3.2
   

   This will download LLAMA 3.2 LLM.
3. Run in terminal:

       pip install -r requirements.txt
   

   This will install the application dependencies.
4. Create .env file with the following variables

       OPEN_AI_API_KEY=<insert OpenAPI key here>
       LANGCHAIN_TRACING_V2=<insert Langsmith key here>
       LANGCHAIN_ENDPOINT=<insert Langsmith key here>
       LANGCHAIN_API_KEY=<insert Langsmith key here>
       LANGCHAIN_PROJECT=<insert Langsmith key here>
   

5. Run the web scraper & parquet reader in the terminal to capture the latest MSME financial information in the web and download the customer loan information provided

       python jobs/load_to_chroma.py
   

System Architecture & Design

How to run the application?

    streamlit run app.py

llm_video.mov

Auditing the LLM results

You can view the logs in Langsmith to backtrack the LLM responses and retrieved data

Contact us

Chester Garett Calingacion
Andrea Aviles
Angela Samboa

Thanks to

Eskwelabs
BPI