LangChain.js & Generative AI

Before using Artificial Intelligence in larger applications, I wanted to better understand what actually happens between a question sent by a user and the answer displayed on the screen.

The LangChain.js & Generative AI Projects repository brings together experiments created during that learning process.

Going beyond the prompt

Using a language model may seem simple: we send a question and receive an answer.

Real applications quickly make that flow more complex.

How can AI access information it does not already know? How can a conversation remember what was said earlier? How can the correct information be found inside a large set of documents?

Those were some of the questions explored in this project.

Exploring different possibilities

Instead of building a single product, I used this repository as a laboratory.

It allowed me to experiment with different models, ways of storing information, and approaches for providing context before an answer is produced.

That was important because it helped me understand the advantages and limitations of each strategy before using them in larger projects.

From the lab to real applications

Part of what I learned through these experiments later influenced other projects.

Concepts related to information retrieval, context, classification, and language model integration appear in a more mature form in applications such as the mAIo Assistant.

For that reason, I mainly see this project as a record of the transition between studying Artificial Intelligence and starting to use it as a software engineering tool.