How to Use AI Reasoning Models: Practical Applications with Hands-On Exercises

Course details

This course explores OpenAI's reasoning model (o-Series) focusing on its advanced reasoning capabilities and practical applications in business and scientific contexts. Instructor Anurag Karuparti covers everything from fundamental concepts and technical implementations to ethical considerations and future development. Test out your new skills along the way with practical hands-on exercises in coding, risk analysis, and safety implementations. By the end of this course, you’ll be prepared to leverage o-Series unique features, including its inference scaling architecture and chain-of-thought reasoning capabilities.

Learning objectives
Explain the key differences between o-Series and previous OpenAI models.
Demonstrate proficiency in crafting effective prompts for o-Series to generate desired outputs across five different scenarios.
Successfully complete a project using o1, o1-mini, o3-mini, demonstrating its advanced capabilities within a one-hour hands-on session.

Concepts

0. Introduction

01 - Introduction to the course
02 - What you should know
03 - Acknowledgements

1. Overview of OpenAI o-Series

04 - What are the o-Series models How are they different
05 - From System-1 to System-2 thinking models
06 - Training scaling laws
07 - The paradigm shift in AI with inference scaling

2. Advanced Reasoning Capabilities

08 - Chain-of-thought reasoning
09 - Majority voting at test-time
10 - Benchmark comparison - Model performance in math, coding, and science
11 - Abstract reasoning

3. Practical Applications and Hands-On Exercises

12 - API features for developers
13 - Best practices for prompt engineering with o-series
14 - How to set up the lab files
15 - Lab 1 - Prompt engineering with reasoning models
16 - Software development
17 - Lab 2 - Game development with reasoning models
18 - Document risk analysis
19 - Lab 3 - Fraud detection with reasoning models
20 - Constraints satisfaction problem
21 - Lab 4 - Employee scheduling with reasoning models
22 - Visual reasoning
23 - Lab 5.1 - Complex floor plan analysis with reasoning models lab 5.2 ERD analysis, SQL generation, synthetic data generation
24 - Evaluation and benchmarking
25 - Understanding and controlling costs

4.Understanding Safety and Security

26 - Security threats and limitations
27 - Building a safer and responsible AI solution

5. The Next Frontier - What Lies Ahead with Reasoning Models

28 - DeepSeek - A powerful open source alternative
29 - Lab 6 - Working with DeepSeek
30 - GPT-5 and beyond
31 - Future outlook

Conclusion

32 - Conclusion