Recently I started learning Generative AI. To be honest, after a very long time I found a technology that really excited me.
As most of us, I started using this technology as my personal assistant (using ChatGPT, Bing Chat etc.) to get answers to the questions that would normally require me to do multiple conventional searches but as I started learning more about it, I realized that its not even the tip of the iceberg. There are so many things you can do with this technology.
In my short course of learning so far, one thing in particular that got me really excited is Retrieval Augmented Generation (RAG). In short, RAG enables you to build Large Language Model (LLM) apps using your own data. Another thing I would like to learn more about are LLM Agents.
Generative AI Learning Path
First one was Generative AI learning course from Google. It is currently available free of cost and is quite good to understand about Generative AI. Without getting too technical, it talks about different aspects of Generative AI and LLMs.
You can access this course here: https://www.cloudskillsboost.google/journeys/118.
Short Courses from DeepLearning.ai
There are a number of short courses available on DeepLearning.ai site that I found incredibly helpful. I did not watch all the courses there but I would recommend watching courses on Prompt Engineering (2 courses), LangChain (2 courses) and Vector Store.
Before taking the Prompt Engineering courses, I used to think why people are making such a big deal about Prompt Engineering but after watching the courses I realized why it is so.
Vector Store course (titled Large Language Models with Semantic Search) explained very nicely what vector stores are and how they work. If you are building RAG based LLM apps, it is really important to understand vector stores and this course does an excellent job.
You can access these courses here: https://www.deeplearning.ai/short-courses/.
LangChain
To test my learning I built a simple app which provided the ability to do semantic search over Azure Documentation and for that I used LangChain. LangChain can be thought of as a very powerful SDK which provides a unified interface to work with multiple LLMs, Vector Stores etc. It also brings in additional features like caching, memory etc. which are not natively available in LLMs.
Though I watched a number of videos however I found that the LangChain courses on DeepLearning.ai to be the best and they have really comprehensive documentation.
You can read the documentation for LangChain here: https://python.langchain.com/docs/get_started/introduction.html.
Generative AI with Large Language Models course on Coursera
I took Generative AI with Large Language Models course on Coursera. I found the course very useful but somewhat overwhelming. I would not recommend taking this course if you are just starting with Generative AI but is worth taking the course once you have had some knowledge in this field.
You can access this course here: https://www.coursera.org/learn/generative-ai-with-llms.
Prompt Engineering for ChatGPT course on Coursera
I took Prompt Engineering for ChatGPT course on Coursera as well. It is completely non-technical and teaches about prompt engineering and various techniques to write effective prompts. I would highly recommend this course.
You can access this course here: https://www.coursera.org/learn/prompt-engineering.
That’s it for this post. I’ll share more as I learn more. Please share what all resources you have been using to learn this incredible technology.
Cheers!
