Building Your First AI Assistant with Large Language Models

Installing the necesary requirements for the project.

Setting up OpenAI API key.

Introduction to what will be covered in this module.

Fast API backbone set up.

Creating first version of an OpenAI Assistant.

Explanations of overall objects encountered when using the Assistants API.

Learn how to use OpenAI Assistant's API and integrate it into your application.

Creating threads with the Assistant's API.

Setting up the Assistant's API endpoint.

Using the OpenAIAssistantRunnable to respond queries from users.

Adding the endpoint to the routes of the application.

Creating module for testing the Assistant through the terminal.

Fixing minor bugs before testing the assistant chatbot.

Chatting with the Assistant through the terminal and testing out it's performance.

Introduction to creating the assistant via code.

Creating the module and changing directory structure.

Setting up the create assistant function.

Explaining the create assistant documentation provided by OpenAI.

Adding the create assistant step to the function.

Fixing minor bugs and testing out the completed assistant.

Comparision between using Langchain vs Open AI Asssistants and a descritpion of what we are going to be doing next.

Installing Jupyter Notebook to the environment.

Loading the portfolio document by using LangChain loaders.

Splitting the document into multiple chunks.

Creating vector embeddings and storing the information within a vector store.

Retrieve relevant documents from the vector store using a query.

Explaining the retrieval augmented generation pipeline that is going to be implemented.

Creating our first RAG pipeline based on the RetrievalQA chain provided by LangChain.

Explaining what LangSmith is and setting it up for monitoring the project.

Creating the RAG with Memory Chain. We'll use LCEL (LangChain Expression Language) to build the pipeline.

Setting up the required components for managing the chat history and the retriever of the chain.

Connecting the pipeline together into a single chain.

Testing the chain and it's ability to interact with the chat history.

We use LangSmith to monitor the performance of the chain while understanding it's inner components.

Streaming the information being generated by the chain.

Integrating the vector store and retriever within the FastAPI application.

Integrating the RAG with memory chain within the FastAPI application.

Introducing the LangServe library for creating simple APIs, docs, and a playground for LangChain apps.

Introducing the LangServe documentation and playground that was automatically generated.

Introdution to the diagram used for deploying the application.

Create the docker image required to containerize the application.

Change to the port through which the appliaction listens and checking if the execution environment is within AWS.

Creating the AWS Serverless Application Model template.

Creating an IAM User in AWS and setting up its required permissions.

Creating an access key for local code.

Virtual Box installation for using AWS within Ubuntu.

Creating Ubuntu 18.04.6 virtual machine.

Fixing VM Ubuntu Bug for opening the terminal.

Installing Virtual Guest Additions to the virtual machine created.

Installation of vscode within the Ubuntu VM.

Copy the project within the Ubuntu VM.

Fix sudo error if account is not in sudoers file.

Installation of latest version of the AWS CLI.

Configure the AWS CLI by using the AWS Access Key ID and the AWS Secret Access Key.

Installing docker in the Virtual Machine. Use the following link:

https://docs.docker.com/engine/install/ubuntu/

Using the SAM build command to build the application inside a docker container.

Using the SAM deployment for the first time

Debugging the error encountered by using Cloud Formation stacks.

Adding the missing policies for our application to be deployed by using Cloud Formation and the SAM template.

Deleting the previous stack used that produced an error.

Using the SAM deploy command to create the AWS Lambda Function.

Testing the deployed application through the LangServe playground.

Explaining how to debug what might be happening with your AWS Lambda application in case you encounter some error while testing it.

Monitoring the applicaiton by using LangSmith once it has been deployed.

Create a simple client from within a jupyter notebook to test out the application.

How to delete and update the application by using SAM CLI.