Session 1 Slides: The AI-Native Workspace

--:-- --
↓ Scroll for more

Session 1

The AI-Native Workspace

AI Product Engineering

Building Intelligent Applications with Modern AI

Block 1: Foundations & The Exocortex | 2 Hours

Today’s Agenda

  • The Old World
  • The New World
  • Three Levels of AI Interaction
  • THINK → KNOW → BUILD → CONNECT → EQUIP → EVALUATE → SHIP
  • Cloud AI Glossary
  • Interactive: Which Level? (3 min)
  • Interactive: Where Does It Belong? (3–4 min)
  • Lab: Set Up Your Workspace
  • Lab: Define Your AI Product Idea

The Old World

Building AI required expensive GPUs, specialised teams, and years of training.

  • Computer vision project: weeks of GPU time + PhD team
  • Recommendation engine: data scientists + ML engineers + infra
  • Chatbot: NLP experts + labelled training data

That world no longer exists for most practitioners.

The New World

Foundation models are accessible via a simple API call.

Send text (or image, audio, document)
Model runs on infrastructure you never touch
Receive sophisticated, structured response
Pay per use — not for the hardware

Three Levels of AI Interaction

LevelExampleResult
Consumer"Give me a recipe"One answer, conversation ends
Prompt Writer"Act as a chef and…"Better answer, still one-shot
Engineering PartnerIterative: define → generate → evaluate → revise → testWorking software

This course operates at Level 3.

Interactive: Which Level? (1/3) (3 min)

True or false — cold-call, then tap.

Interactive: Which Level? (2/3) (3 min)

True or false — cold-call, then tap.

Interactive: Which Level? (3/3) (3 min)

True or false — cold-call, then tap.

THINK → KNOW → BUILD → CONNECT → EQUIP → EVALUATE → SHIP

🧠 THINK — Explore problems with AI

📚 KNOW — Build your knowledge base

🔨 BUILD — Scaffold with AI coding partner

🔌 CONNECT — Wire to cloud AI APIs

🛠️ EQUIP — Add tools & agentic workflows

🧪 EVALUATE — Test for failures

🚀 SHIP — Deliver working prototype

Interactive: Where Does It Belong? (3–4 min)

Classify each activity into the course framework stage.

Cloud AI Glossary

TermMeaning
Foundation modelPre-trained general capability
APIStructured request/response to the model
TokenUnit of text (cost + context)
InferenceRunning the model for a response
Context windowHow much text the model sees at once

Lab: Set Up Your Workspace (1/2)

  1. GitHub account + Copilot (Education Pack / trial)
  2. VS Code + Copilot extension

Lab: Set Up Your Workspace (2/2)

  1. Node 20+ — node -v
  2. AI Studio → freeform prompt → get API key (store safely)
  3. Verify: open .js, type function get — Copilot suggests

Lab: Define Your AI Product Idea

One project for all 15 sessions. Answer all four:

QuestionStrong answer sounds like…
ProblemSpecific waste of time / failure mode
UserNamed cohort (not "anyone")
AI RoleWhat the model actually does
SuccessMeasurable (e.g. 8/10 grounded answers)

Checkpoint: 60-second pair pitch — partner names one weakness.

Session 1 Summary

  • Cloud AI = foundation model capability via API call
  • AI as engineering partner, not search engine
  • Framework: THINK → KNOW → BUILD → CONNECT → EQUIP → EVALUATE → SHIP
  • You have a product idea, a workspace, and an API key

Next Session: Prompting & Structured Outputs