Phase 2

• Duration: Six Sessions
• Sessions: One per week (60 minutes each)
• Learning Style: Hands-on, project-based, and data-driven
• Capstone Project: Build an AI-powered reasoning engine that predicts, analyzes, and improves over time
• Best For: Students, data enthusiasts, and future AI engineers looking to understand AI decision-making

This course is structured into three phases:

• Weeks 1-2: Understanding AI Reasoning and Predictions – How AI thinks, analyzes, and learns from data
• Weeks 3-4: Building AI Reasoning Models – Training AI on real-world datasets for decision-making
• Weeks 5-6: Optimizing, Testing, and Deploying – Making AI smarter, refining accuracy, and final showcase

Goal: Learn how AI analyzes data, recognizes patterns, and makes predictions.

Tools Used:

• Google Colab 
• Perplexity 

Key Topics: 

• How AI predicts the future by analyzing past data.
• Examples of AI-powered recommendations and fraud detection.

Hands-On Activity:

• Students compete against an AI model to predict trends (sports, weather, stock prices).
• Students load a real dataset into Google Colab and train an AI model to recognize simple patterns (e.g., temperature trends, sales forecasts).

Outcome: Students understand how AI models analyze data and predict outcomes.

Goal: Learn how AI uses training data to improve predictions.

Tools Used:

• Google Colab
• Hugging Face

Key Topics:

• How AI learns from mistakes and improves with more data.
• Examples of AI predicting housing prices based on multiple factors.

Hands-On Activity:

• Students predict an event outcome before comparing it to an AI-generated prediction.
• Students fine-tune an AI model using Hugging Face’s dataset, tweak training parameters, and test results.

Outcome: Students will build an AI model that continuously improves based on performance data.

Goal: Learn how AI uses multiple variables to make complex decisions.

Tools Used:

• Google Colab
• OpenAI API

Key Topics:

• How AI considers multiple factors when making decisions.
• Examples of AI assisting in medical diagnoses and investment recommendations.

Hands-On Activity: 

• Students compare their decision-making logic with AI-powered recommendations.
• Students train an AI model to analyze a dataset and recommend the best choice.

Outcome: Students build AI models that analyze data and provide decisions.

Goal: Train AI to adjust decisions based on new data and feedback.

Tools Used:

• Hugging Face
• OpenAI API

Key Topics:

• How AI improves based on feedback.
• Examples of Google Translate improving over time.

Hands-On Activity: Students test an AI chatbot and analyze incorrect responses.

Students train an AI model and refine responses based on feedback.

Outcome: Students build AI models that improve over time.

Goal: Explore real-world AI applications and ethical concerns

Tools Used:

• Google Colab 
• Perplexity

Key Topics:

• The role of data in AI decision-making.
• How biased data can lead to unfair AI outcomes.

Hands-On Activity: 

• Students research and present real-world AI failures (biased hiring, fake news algorithms).
• Students analyze a flawed AI model and propose solutions.

Outcome: Students learn how to make AI fair and trustworthy.

Goal: Apply everything learned to build a real AI-powered reasoning engine.

Final Capstone Project: Students create an AI model that predicts, analyzes, and provides recommendations.

Hands-On Activity: Students pitch their AI models and showcase their projects to mentors and peers.

Outcome: Students develop a fully functional AI reasoning model.