Hello Learners…
Welcome to the blog…
Table of Contents
- Introduction
- Skills You Need to Become a Generative AI Engineer
- Prerequisite For Generative AI
- Important skills for generative ai
- Summary
- References
Introduction
In this post we discuss the Skills You Need to Become a Generative AI Engineer,if you are really interested in understanding about generative ai work and you are really want to work in this field then this blog post is for you.
Skills You Need to Become a Generative AI Engineer
Whenever we talk about generative ai engineering we really want to talk about two important thing one is the prerequisites and the second thing is that what are the important skillset.
- Prerequisite for generative ai
- Important skills for generative ai
Prerequisite For Generative AI
- Python
- Machine Learning
- Deep Learning
- Computer Vision
- NLP Concepts
- Transformers
Important Skills For Generative AI
In Generative AI we probably work with three types of large language models,
- LLM (Large Language Model)
- LIM (Large Image Model)
- Multi Model
LLM (Large Language Model)
Large language model used when our use case is related to text data, like chatbots ,text summarization etc.
LIM (Large Image Model)
Large Image model used when our use case is related to image data, like image generation etc
Multi Model
Multimodel is a combination of text and images so that types os model is used to solve use cases related to both text data and image data.
Example: Google gemini pro , ChatGPT 4
In Generative AI our main aim of the model us to generate new content based on any context it will able to generate the required data.
Skillset Required For Generative AI
Right now there are open source and paid Generative AI models availabe, and we have to know about both and try to understand that models and how they works and also how we can use that models in our use cases.
We have to practice with different different models and try to understand how to implement and how to deploy that models in production.
Open Source Models
- LLAMA 2 by Meta
- Mistral
- Falcon
https://galaxyofai.com/tag/open-source/
Pain Models
- OpenAI
- Claude 2
- Mistral
- Google Gemini Pro
We can use both open source and paid models to address our use cases. However, when considering deployment and scalability, we often rely on cloud infrastructure. As a result, many opt for paid models, minimizing the need for extensive cloud computing power.
In case of paid model they have their personal cloud and we can use directly from their cloud throug their APIs.
Also there are services for that on AWS and AZURE.
In AWS there is a service named Bedrock, which has every llm models, stable diffusion models. Bedrock have all the functionalities with respect to the open source models and paid models.
They provide direct an API for finetuning or generate the new content, we don’t have to worry about cloud part.
Library that we can use or learn to use Generative AI Models?
- There is a library called HunggingFace that provide packages for Open Source Generative AI Models using that we can use that models.
Frameworks that we have to learn for Generative AI?
- The most important framework to be aware of is the OpenAI framework.
- Two of the most important frameworks currently in use are LangChain and LLamaIndex
Using langchain and llamaindex we can call both open source and paid generative ai models.
Frontend Frameworks that er have to learn For Generative AI Practice?
For practice generative ai models we can use Streamlit for the frontend part.
Vector Databases that we have to learn For Generative AI?
There are different types of vector databases available like chromadb, cassendra, data stacks. We have to try that and also understand that how the specific types of vectore database is work in production.
The most important thing in all this how we can finetuning generative ai models for our custom data using all the above tools.
Summary
In summary, it is crucial to understand both open source and paid LLM models, along with gaining insights into cloud platforms. This knowledge is essential for effectively navigating and leveraging diverse approaches in the rapidly evolving landscape of Generative AI.