projects by Pablo Elgueta
.jpg)
Build a Custom Chatbot Using LangChain and ChatGPT
InfoHub has a groundbreaking idea—a chatbot that can answer a company’s questions about its knowledge base just like the user was conversing with a real human expert! This idea can now become a reality thanks to the latest advances in large language models (LLMs). That’s exactly what you’ll be doing in this liveProject series! You’ll step into the developer role at InfoHub, and deliver a chatbot capable of everything from troubleshooting tech issues to answering queries about product specifications. You’ll integrate GPT-3.5 into a chatbot to understand user questions and retrieve related embeddings, utilize ChromaDB to manage storage and retrieval, and then deploy the whole system via Streamlit.
Build and Deploy the Chatbot
In this liveProject, you’ll step into the role of a software developer at InfoHub, an AI startup attempting to build a chatbot that can answer queries about a company’s knowledge base. You’ve already built the basics of the chatbot—now it needs to be deployed! To do this, you’ll use the LangChain feature LangServe to create a LangChain server application. You’ll build your chatbot using Streamlit and then deploy it to the Streamlit cloud.
Q&A Using Vector Databases
In this liveProject, you'll join InfoHub, an AI startup disrupting corporate knowledge management. They aim to unlock a company’s knowledge base through conversational Q&A-style user interfaces that use breakthrough language models. You'll leverage LangChain, a framework optimized for integrating LLMs into apps, to integrate InfoHub's data, vector stores, and language models into a single solution. You’ll prepare your data, create a vector store to embed your documents, and then use LangChain to combine it with an LLM.
Augment with Embeddings
In this liveProject, you’ll become a software engineer at InfoHub, an up-and-coming AI startup looking to revolutionize how companies interact with their knowledge bases. InfoHub seeks to utilize groundbreaking large language models to deliver a system whereby a user’s questions about company data can be answered through a Q&A-style language interface. You’ll begin by assisting them in creating this tool by processing, tokenizing, and converting data into embeddings with BeautifulSoup and tiktoken.
