AI in eCommerce​

1. Google Teachable Machine: Build a no‑code image or audio classifier from sample data.
2. Scikit‑Learn Demo in Colab: Run a simple classification example (e.g., Iris dataset).
3. Deliverable: “Train Your First Model” write‑up—export your Teachable Machine model, report its accuracy metrics, and describe one potential application (1 page).

Tools & Activities (2 hours hands‑on):

• Google Teachable Machine: Build a no‑code image or audio classifier from sample data.
• Scikit‑Learn Demo in Colab: Run a simple classification example (e.g., Iris dataset).
• Deliverable: “Train Your First Model” write‑up—export your Teachable Machine model, report its accuracy metrics, and describe one potential application (1 page).

• TensorFlow Playground: Experiment interactively with hyperparameters to see their effects on a toy dataset.
• Keras in Google Colab: Run a prewritten notebook to train a basic neural network on MNIST (handwritten digit recognition).
• Deliverable: Submit your Colab notebook screenshot showing model architecture, training/validation metrics, and a brief reflection (200–300 words) on parameter tuning and achieved accuracy.

Tools & Activities (2 hours hands‑on):

• TensorFlow Playground: Experiment interactively with hyperparameters to see their effects on a toy dataset.
• Keras in Google Colab: Run a prewritten notebook to train a basic neural network on MNIST (handwritten digit recognition).
• Deliverable: Submit your Colab notebook screenshot showing model architecture, training/validation metrics, and a brief reflection (200–300 words) on parameter tuning and achieved accuracy.

• Hugging Face Transformers in Colab: Run a text‑classification demo using a pretrained BERT or RoBERTa model on sample data.
• NLTK/spaCy Demo: Perform simple sentiment analysis or entity recognition on a small text corpus.
• ChatGPT Prompt Workshop: Write and test prompts for automating tasks like “Generate a customer‑support email” or “Summarize this article in bullet points.”
• Deliverable: Provide (1) two Hugging Face Colab notebook screenshots (text classification and sentiment analysis) and (2) a list of five tested ChatGPT prompts with before/after outputs and a short rationale (200–300 words total).

Tools & Activities (2 hours hands‑on):

• Hugging Face Transformers in Colab: Run a text‑classification demo using a pretrained BERT or RoBERTa model on sample data.
• NLTK/spaCy Demo: Perform simple sentiment analysis or entity recognition on a small text corpus.
• ChatGPT Prompt Workshop: Write and test prompts for automating tasks like “Generate a customer‑support email” or “Summarize this article in bullet points.”
• Deliverable: Provide (1) two Hugging Face Colab notebook screenshots (text classification and sentiment analysis) and (2) a list of five tested ChatGPT prompts with before/after outputs and a short rationale (200–300 words total).

• Power BI with AI Visuals: Connect to a sample sales dataset, generate AI‑powered insights, and create a brief automated report.
• MonkeyLearn Workflow: Upload customer reviews and train a no‑code text classifier for sentiment tagging.
• ChatGPT/Copilot Demo: Generate a business report summary or automate a standard email template.
• Ethics Debate & Reflection: Participate in a guided discussion—“Should AI replace human hiring decisions?”—and complete a short reflection quiz in Google Forms.