Upwork is hiring a Gen AI Vertex AI Langchain Streamlit App

I am looking to create a streamlit ( UI ) app - chatbot that can take input as natural language query, on submission of the query, it will execute python code that leverages SQLalchemy ( to create bindings to the table in Bigquery ), send the query to vertex AI LLM ( text-bison ), generate the SQL, execute the query against the dataset and bring back the result in the streamlit app. If the result set is more than 1 row, for eg: what are my top 5 products or who are my top 10 customers, it will also generate a chart/graphical representation. The chatbot should hold the memory of the conversation when answering follow up questions: who is my top customer - answer: ABC, next question - what are the top products I sell to this customer: Answer: x,y,z - in this scenario it should understand that the customer in this ask is ABC.