Please see the attached for the instructions. I am working on a project, which i

Please see the attached for the instructions.
I am working on a project, which is to build a food recommendation system using an LLM for diabetic patients. Specifically, I have a CSV dataset of the following columns:
1) Name of Food Centre
2) Name of stall (each food centre has many stalls)
3) Address
4) Cuisine
5) Various categorical variables of nutritional aspects (e.g. Low in sugar, lower in calories, etc.)
6) Food dish
Currently, what I’ve done is:
1) Build a knowledge graph out of the dataset
2) Use a RAG-based approach to query the dataset with GPT-3.5 via prompt-templating
Moving forward, I would like to work on the following tasks:
1) Memory
2) Agents (e.g. calorie calculation where diabetic patients can enter their information and we calculate their calorie intake, google map for users to find the exact location of the food stall)
3) Multi-modal (Users can take picture of the food and the LLM can perform some calculation e.g. amount of calories)