Getting ready to build out Agentic Apps in 2025? Check out these articles.
Article Updates
- January 1, 2025: First version
- January 6, 2025 : Added Agents Whitepaper by Google and HuggingFace’s Introduction to Agents
As a builder, one of the emerging themes over the 2nd half of 2024 has been Agents and how they are all set to change the way we work by assisting us with our tasks and possibly taking the next step too of actionable steps that they perform on their own with minimal to no supervision.
There are several frameworks of note that help you build out Agents, both single and multi-agent architectures and with the ability to act or wait for the human-in-the-loop response. At a high level, you would have even read that Agents are composed of key building blocks : The brain (model), memory and a set of tools.
This blog post highlights a few key articles that have helped me understand this space better. This is essential because it allows you to then go an evaluate the Agent frameworks with a lot more of conceptual grounding and even evaluate if they support a certain feature or not.
Here is a list (in no order of preference), which I have highlighted on LinkedIn too in a post:
1. Sanjeev Mohan’s interestingly titled article “Demystifying AI Agents: Frequently Asked Questions (FAQ)”. A fantastic deep dive FAQ that is easy to read and breaks down the key buildings blocks well.
2. Anthropic’s blog post “Building Effective Agents” covers some key points and helps to understand Workflows v/s Agents. Personally, this article helped me differentiate between these two points well. What I thought were Agentic Apps that I was building are actually Workflows as per this paper. And I liked the distinction that it made.
3. Swapan Rajdev’s “Understanding Agentic AI Architecture” is a solid read on Agentic Architectures, Frameworks, the current ecosystem and practical consideratons.
4. Janakiram MSV has probably one of the most detailed part by part tutorial on AI Agents. It starts off with understanding what AI Agents mean from a Developer perspective and then builds on that by demonstrating how to build out an Agent in subsequent parts, right from persona, tools, memory and more.
5. Dr. Sokratis Kartakis article titled “Demystifying Generative AI Agents” is a solid and no-nonsense article that defines what an Agent is and then intersperses that with code snippets to help to immediately place it in content on how you would implement one.
6. Agents Whitepaper by Google highlights the fundamental blocks of an Agent, how to connect them to the outside world and Langchain examples for Agents.
7. Lavi Nigam takes things a step further with his notebook that demonstrates how you would implement an Agent, that implements Gemini Advanced Deep Search online research feature. He implements a MultiAgent system design pattern using Gemini 2.0 on VertexAI. This is a detailed notebook and there is a lot to learn from this, if you are looking to build one out on your own, without relying too much on any specific framework.
8. HuggingFace has authored an article titled “Introduction to Agents”, which is part of its Smolagents documentation. What particularly stood out here are two things: “level of agency” definition i.e. the amount of independent work that an Agent can do and their concept of CodeAgent, that puts code execution at the center of the reasoning and act process to execute tasks.
Happy Reading ! I request you to share any similar articles that helped you understand this space too. Please do so in the comments.
