Reasoning & Architectures

Hands-on lab with starter code you can copy + run, and guided steps to produce the artifact for this module.

Setup

Open a fresh terminal in a new folder, e.g. mkdir lab-2 && cd lab-2.
Create a Python 3.11 venv: python -m venv .venv && source .venv/bin/activate.
Install the dependencies noted on each starter snippet (the small pip install … line above the code).
Copy each starter into a file with the suggested name, then run it and capture the output.

Starter code

01 · ReAct loop — interleaved Thought / Action / Observation

save as: 02_01_lab.py

python

import re, json
from openai import OpenAI
client = OpenAI()

TOOLS = {"calc": lambda x: str(eval(x, {"__builtins__":{}}))}

SYS = """Use the format:
Thought: ...
Action: tool_name[input]
Observation: ...
... (repeat)
Final Answer: ..."""

def react(question, max_iter=6):
    history = f"Question: {question}\n"
    for _ in range(max_iter):
        out = client.chat.completions.create(
            model="gpt-5-mini",
            messages=[{"role":"system","content":SYS},
                      {"role":"user","content":history}],
            stop=["Observation:"],
        ).choices[0].message.content
        history += out
        m = re.search(r"Action: (\w+)\[(.+?)\]", out)
        if not m:                       # Final Answer reached
            return history
        tool, arg = m.group(1), m.group(2)
        obs = TOOLS[tool](arg) if tool in TOOLS else f"unknown tool {tool}"
        history += f"\nObservation: {obs}\n"
    return history

print(react("What is 17 * 23 + 4?"))

Run with: python 02_01_lab.py

Guided steps

Work through each step in order. Stop when your output matches the expected artifact — then push your work to your lab folder.

1
ReAct calculator
Do — Implement the ReAct loop with a calc tool. Test on 'What is 17*23+4?' and 'sqrt(144)+2^8'.
Expected artifact — Transcript showing Thought / Action / Observation steps and Final Answer: 395 and 268.
Run free:
2
Plan-and-execute variant
Do — Replace ReAct with a Plan-then-Execute architecture (LLM emits a 3-step plan, executor runs each).
Expected artifact — Same 2 questions answered correctly; transcript shows the explicit plan before any action.
Run free:
3
Architecture comparison
Do — Run both architectures on 20 mixed math/reasoning prompts. Report accuracy and avg steps.
Expected artifact — arch.md with a 2×2 table (acc, steps) and a 3-sentence recommendation.
Run free:

Authoritative sources

Independent primary sources used to check and corroborate the guidance on this page.

NIST National Initiative for Cybersecurity Education (NICE)
US national framework for cybersecurity work roles and competencies.
Financial Action Task Force (FATF)
Global AML/CFT standards referenced by financial-crime curricula.
FBI Internet Crime Complaint Center — Annual Reports
Primary loss and typology data used in fraud-investigation training.
US Federal Trade Commission — Consumer Sentinel Data Book
Official complaint statistics for teaching fraud trend analysis.