Practical Deep Agents: Implementing Autonomous Problem-Solving Loops

Explore the evolution of AI development from standard large language models (LLMs) to the emerging field of "deep agents," also known as deep research agents. Gain a foundational understanding of how basic AI agents—referred to as "shallow agents"—operate by using an LLM as a "brain" to interact with external tools and APIs. Learn about the significant limitations of LLMs in planning, memory, and handling complex queries. Review conceptual explanations and code examples to learn how complex queries can be decomposed into manageable sub-queries. By the end of this course, you'll be more prepared for practical implementation of advanced, autonomous AI solutions.

Learning objectives
Describe the evolution of AI development from LLMs to independent agents, multi-agent collaboration, and deep agents.
Define "deep agents" (also known as deep research agents) and describe how these are different from "shallow agents".
List components of basic agents, including how an LLM composes output and interacts with external tools like weather or search APIs.
Describe the limitations of shallow agents, such as their inability to perform explicit planning, limited context retention, and difficulty handling complex, multi-part queries.
Explain how complex queries are decomposed into sub-queries to be solved.
Identify code examples for creating deep agents and describe the foundational concepts for practical implementation.