An AI agent is a software program that can interact with its environment, collect data, and applies AI on the data to perform certain tasks. Regardless of the goal set by humans, an AI agent decides the best actions to achieve those goals.
Roadmap to learning to build your first AI agent
As an example, if you are on your hospital website and are chatting with an AI agent. The AI agent will ask the patient several relevant questions and verify the answers in internal documents. Based on that, it will respond with a solution. Depending on what the patient replies, the AI agent then determines whether it can resolve their issue or get a human involved.
This is a brief high level description of steps needed to learn how to program your own AI agent.
Basics of Gen AI and RAG: Introduction to Generative AI
Generative AI or GenAI is a type of AI that can create new content like stories or articles, or generate media like images, videos, and music.
It can learn about different topics like recipes, art, programming, math, art, science, etc, and use its newfound knowledge to solve problems or aid in recommendations.
As an example, someone writing a 1,000 word essay can use Gen AI to suggestions to overcome the writer's block. Similarly, businesses can use Gen AI in many ways, including building chatbots, creating thumbnails, posters, designing prototypes and products, and even ideas for future work.
Recently, high schoolers in Chicago created a website that uses Gen AI to design clothes for customers to buy.
You can get started with Introduction to Generative AI from Google Cloud Skills Boost
You can also watch the video from Google Cloud Tech.
Basics of Prompt Engineering
The next thing on the list are the fundamentals of Prompt Engineering.
A Prompt consists of instructions and context passed to a language model to achieve a task and get corresponding responses.
Prompt Engineering is the process of using and structuring input prompts to get more refined and relevant responses from the language model.
This is a good prompt engineering video tutorial on YouTube.
Data Processing
Data Processing is the method of converting raw data into useful relevant information. This is useful for various domains including business, finance, medical health and science.
In data processing, we define and manage the structure, characteristics and specifications of of data within the entity that owns or maanages it.
It can involve sorting, searching through the data and converting it to a different format.
The steps in data processing involve Collection, Preparation, Input of Data, Data Processing, Data Output in a presentable and desired format and finally Data Storage.
Google has a video about data processing in their Data Analytics Certificate video.
Basics of RAG
Retrieval Augmented Generation, or RAG, provides LLMs with the ability to retrieve information from one or multiple data sources and use that information to answer a query.
In basic RAG, this is the general workflow.
Query -> Query Embedding -> Similarity Search -> Retrieval -> Context -> Answer
There are many tutorials out there on building a RAG from scratch. This blog post appears to be straightforward for beginners.
You may also watch this video on how to create a RAG from scratch.
Basics of API Wrappers
An API wrapper is a layer of software that sits between an application and an existing API. This allows the application to interact with the API in a standardized way. API wrappers provide a simplistic interface for developers to access and manipulate data from various APIs, leveraging APIs from OpenAI, Gemini, etc.
Building AI agents around API wrappers will make it efficient. You get improved data processing, you can do more efficient decision making, you can automate repetitive tasks and integrate data better from multiple data sources using their APIs.
You can watch this video.
Introduction to AI Agents
An AI Agent is a software program that is designed to perform specific tasks or make decisions on behalf of a user or organization. An AI agent can be trained to learn from data, recognize patterns or make predictions / recommendations. AI agents can be used to enhance the functionality of AI wrappers by providing intelligent decision-making capabilities.
Here's a video explaining AI agents.
Agent Frameworks
Agent Frameworks are software systems that enable the creation of autonomous agents capable of making decisions and taking actions based on user inputs, data, or pre-defined rules.
These frameworks allow the development of agents that can understand natural language, interpret complex instructions, and perform various tasks automatically. They leverage AI models (LLMs) to interpret prompts and execute actions like API calls, database queries or automation.
You can watch a video explaining agentic frameworks in detail.
Building our First AI Agent
This YouTube video is good for learning how to build our first AI agent in Python.
This is another video that is useful.
Conclusion
This is just a roadmap. Your progress depends on your current knowledge of the area, both technical and theory. Do not rush into learning it. At the same time, if you have an idea, do not wait to make it perfect before giving it life.
I will post separate blog posts for AI-related topics regularly. Thank you for reading. Suggestions and comments are always welcome!
Related Posts
If you have any questions, please contact me at arulbOsutkNiqlzziyties@gNqmaizl.bkcom. You can also post questions in our Facebook group. Thank you.