Developing LLM Applications with LangChain
Build chatbots, tool-using agents and RAG pipelines in Python with LangChain, using OpenAI and Hugging Face models and a vector database.
Large language models only know what was in their training data, and on their own they cannot reach outside data or call external functions. LangChain is a framework that links models, prompt templates and vector databases through shared building blocks. You build chatbots on open-source and proprietary models, pass inputs through multi-step chains, let an agent pick which tool to call, and ground a model’s answers in text loaded from PDF, CSV and HTML files.
What You Will Achieve
- Connect OpenAI and Hugging Face models to a chatbot and steer its replies with a set of few-shot examples.
- Pass one model’s output into the next step of an application using the LangChain Expression Language.
- Write a Python function, wrap it as a custom tool, and let a ReAct agent decide when to call it.
- Load and split documents, then answer questions with a RAG chain that retrieves passages from a vector database.
Program Overview
LangChain Components and Prompt Templates
Sets out the main parts of the framework and builds chatbots on open-source and proprietary models with increasingly structured prompts.
- Models, chains, agents, prompts and parsers
- OpenAI and Hugging Face models
- Chat prompt templates
- Few-shot prompting and example sets
Sequential Chains and Agents
Moves to multi-step applications in which inputs flow between components and a model chooses which action to take next.
- Sequential chains with LCEL
- ReAct agents
- Defining functions as custom tools
- Integrating tools with agents
Retrieval Augmented Generation with Vector Databases
Addresses the training cut-off of LLMs by bringing outside documents into the application through the full retrieval workflow.
- PDF, CSV and HTML document loaders
- Character, recursive and HTML splitting
- Storing embeddings in a vector database
- Retrieval prompt templates and RAG chains
Certificate Awarded (Paid)
Job Outlook
These skills apply wherever teams build chatbots, document question-answering assistants or other user-facing tools on top of large language models, mostly in software engineering and AI application development.
Good for: Developers who want to add chatbots, tool-using agents and question answering over their own documents to Python applications.
Keep in mind: The three chapters center on prompts, chains, agents and the retrieval workflow; debugging, performance evaluation and chatbot memory get no lesson of their own.
Bottom line: A code-first look at LangChain whose retrieval chapter walks through each step of turning your own files into a working RAG chain.
Frequently Asked Questions
What is the difference between a chain and an agent in LangChain?
A chain runs a fixed sequence of components, such as a prompt template followed by a model. An agent lets the model decide which actions to take, including calling external functions or tools. The course builds both, starting with chains.
Which document formats does the retrieval chapter work with?
The RAG chapter loads PDF, CSV and HTML files, splits them by character, recursively or by HTML structure, and stores their embeddings in a vector database so the most relevant passages can be retrieved.
Does the course use open-source or proprietary models?
It uses both. Early lessons connect proprietary OpenAI models and open-source Hugging Face models to LangChain, so you see how each type is called and prompted through the same framework.


There are no reviews yet.