📘 Prompt Engineering & Context Engineering

Workshop Materials · FACTS-H · IIIT Kottayam

This repository contains all teaching materials for the session on Prompt Engineering, Context Engineering, and Verbalized Sampling (VS).
The resources are designed for students, educators, and practitioners interested in building stronger, context-rich, and responsible AI prompts.

---

📂 Repository Contents

1️⃣ Student Activity Handout

• handout.pdf
• handout.tex
  A mobile-first activity guide that helps students transform weak prompts into strong, context-driven prompts using RTAFCE, Context Engineering, and Verbalized Sampling.

2️⃣ Slide Deck

• Prompt_Engineering_slides.pdf
  Slides covering:
• Why prompting matters
• RTAFCE framework
• Context Engineering
• Verbalized Sampling
• Toolchain recommendations
• Integrated workflow
• Group activity instructions
• Responsible AI guidelines

3️⃣ Supporting Files (if added)

• Any demo prompts
• Example strong/weak prompt lists
• Additional reading references

---

🎯 Learning Objectives

By the end of this session, learners will be able to:

• Understand the RTAFCE structure for strong prompts
• Engineer context to control tone, audience, medium, and purpose
• Use Verbalized Sampling (VS) to increase diversity and creativity
• Apply a toolchain workflow (PromptTessor / Quartzite → RA-iPromptingToolkit → ChatGPT → Perplexity)
• Evaluate prompts for risks such as vagueness, clichés, bias, and platform mismatch
• Produce responsible, platform-aware AI-generated media content

---

🧠 Key Concepts

🔹 RTAFCE Framework

• Role
• Task
• Audience
• Format
• Constraints
• Examples

🔹 Context Engineering

Designing the background, expectations, and medium in which the AI’s response will live.

🔹 Verbalized Sampling (VS)

Prompting the model to:

• explore distributions
• generate multi-style outputs
• reveal low-probability creative ideas
• avoid predictable or “safe” mode-collapse responses

---

📱 Usage (Mobile-First Activity)

Students can complete the entire activity using:

• WhatsApp / Telegram group
• Google Docs / Notepad
• Padlet / online boards

No chart paper or laptops required.
Each group selects a weak prompt and rewrites it using the five-step workflow.

---

🛠 Recommended Tools

• ChatGPT (content generation)
• PromptTessor / Quartzite (RTAFCE scaffolding)
• FACTS-H RA-iPromptingToolkit (prompt comparison)
• Perplexity / Gemini (verification and fact-checking)

https://promptcraft-academy-695890547372.us-west1.run.app/

https://ra-ipromptingtoolkit.vercel.app/

🏫 About FACTS-H

FACTS-H (Fairness, Accountability, Transparency, Social-Humanities Lab) at IIIT Kottayam focuses on responsible, explainable, and human-centered computing.
This workshop align