Building AI Agents: Autonomous Systems with LLMs

AI agents combine LLMs with tools and decision-making logic to autonomously complete complex tasks.

Basic Agent Pattern

from langchain.agents import AgentExecutor, create_react_agent
from langchain.llms import OpenAI
from langchain import hub

# Define tools
from langchain.tools import Tool

def calculator(query):
    return eval(query)

def search_web(query):
    # Call search API
    return f"Search results for: {query}"

tools = [
    Tool(name="Calculator", func=calculator, description="Useful for math"),
    Tool(name="Search", func=search_web, description="Search the web")
]

# Create agent
llm = OpenAI(temperature=0)
prompt = hub.pull("hwchase17/react")

agent = create_react_agent(llm, tools, prompt)
agent_executor = AgentExecutor(agent=agent, tools=tools, verbose=True)

# Run agent
result = agent_executor.invoke({"input": "What is 25 * 4 plus 10?"})

ReAct (Reasoning + Acting)

Thought: I need to calculate this step by step
Action: Calculator
Action Input: 25 * 4
Observation: 100
Thought: Now I need to add 10
Action: Calculator
Action Input: 100 + 10
Observation: 110
Thought: I have the final answer
Final Answer: 110

Custom Tools

from langchain.tools import BaseTool

class DatabaseQueryTool(BaseTool):
    name = "database_query"
    description = "Query the database for information"
    
    def _run(self, query: str) -> str:
        # Execute database query
        results = execute_sql(query)
        return str(results)

class EmailSenderTool(BaseTool):
    name = "send_email"
    description = "Send an email to a recipient"
    
    def _run(self, to: str, subject: str, body: str) -> str:
        # Send email
        send_email_api(to, subject, body)
        return "Email sent successfully"

# Add to agent
tools = [DatabaseQueryTool(), EmailSenderTool()]

Multi-Agent Systems

# Coordinator agent
coordinator = Agent(
    role="Project Manager",
    goal="Coordinate tasks between agents",
    tools=[]
)

# Specialized agents
researcher = Agent(
    role="Researcher",
    goal="Research information",
    tools=[search_tool, scrape_tool]
)

writer = Agent(
    role="Writer", 
    goal="Write content",
    tools=[write_tool, edit_tool]
)

# Execute workflow
task = "Research AI trends and write a blog post"
result = coordinator.delegate(task, [researcher, writer])

AI agents enable autonomous task completion. Build sophisticated systems today!