MC04 is a foundational module inside the AI Residency Program. In this session, we’ll break down how Large Language Models (LLMs) actually work (in practical terms) and then move into a guided, hands-on lab using the OpenAI Playground to learn prompt iteration, parameter tuning, and output control—exactly the skills you’ll use in later modules for RAG, agents, and production-grade AI apps.

What you’ll learn (outcomes)
By the end of this session, you will be able to:

• Explain LLMs in a clear, correct mental model (tokens, context window, embeddings at a high level, training vs inference).
• Understand what prompting really is: instruction hierarchy, constraints, examples, and evaluation.
• Use Playground effectively to:
• Iterate prompts quickly with versioning mindset
• Tune parameters (temperature, max tokens, top-p) based on task type
• Compare outputs and reduce randomness / hallucinations with structure
• Create a reusable prompt template for your own AI Residency project.

Agenda (90–120 mins)

1. LLM fundamentals that matter for builders
2. Why outputs vary: tokens, probability, sampling, context
3. Prompt patterns: role/task/constraints/examples + formatting
4. Playground lab: prompt iterations + parameter tuning
5. Mini-challenge: build a “reliable response” prompt for your project
6. Q&A + what this unlocks in upcoming modules

Who this is for
AI Residency participants building along any track (LLM/API, no-code/low-code, agentic workflows). Ideal if you want a strong base before moving into advanced builds.

Prerequisites

• Basic familiarity with ChatGPT-style prompting is enough
• Add 5$ in OpenAI platform
• No heavy coding required for this module (coding comes in later modules)
  

Joining link

AI Residency Program / Access Portal: https://decodingdatascience.com/airesidency

If you’re already enrolled: the live session join link (Zoom/stream) is typically shared inside the Residency portal and/or your session email/calendar invite.