Hackathon

Ship Day: Studio Sprint

Overview

# Overview

Curriculum.Units

Shipping a working, multi-page product against a deadline in role-based teams. Students research a comparable company, diagram how code, database, AI and APIs connect, and deliver a landing page, a live-data gallery and a grounded agent.

# Objectives

Curriculum.Units
  • 01 By the end, you can Research a real comparable company and use what its site teaches in your own plan.
  • 02 By the end, you can Work to a set team role (CEO, CTO, marketing or CFO) and ship to a fixed deadline.
  • 03 By the end, you can Draw an architecture diagram showing how the code, the database, the AI and the APIs connect.
  • 04 By the end, you can Combine both to ship three working pages — a landing page, a live-database gallery, and a password-protected agent — behind a one-sentence pitch.

# Sessions

Curriculum.Sessions
  1. Team Roles and Kickoff
    Applied craft
    • Worksheet Trace The Dollar Teams list every piece of AI content their product needs, mark which ones cost money, and add up the bill.
    • Game Free Or Paid Drill The teacher shows a task and teams quickly shout whether the AI does it for free or for money.
  2. Mapping the System
    Applied craft
    • Role-play Be The Pipe Learners act out the steps of a request, first waiting frozen for the result, then carrying on while it loads.
    • Simulation Papers Please Stamping Learners play an AI that approves or rejects documents against set rules, with less time as the queue speeds up.
  3. Testing and Shipping
    Deliverable prep
    • Quiz Name The Error The teacher shows an error message and teams race to say what caused it.
    • Game Gandalf Password Defense Learners try to trick an AI into revealing a password as its defences get tougher at each level.
    • Warm-up Attack Your Own Agent Warm-up Teams spend the first minutes trying to break their own agent's defences and list every weakness they find.
  4. Pitch and Premiere
    Assessment
    • Worksheet One-Sentence Pitch Builder Each team fills in a short template that squeezes their product into a single clear sentence saying who it helps and how.
    • Role-play Pitch Rehearsal Rounds Teams pair up and take turns saying their one-sentence pitch out loud, then give each other two things that worked and one thing to fix.
    • Demo Demo Day Pitch-Off Every team delivers its final one-sentence pitch and shows the live site while a panel scores each one against a shared rubric.

Applied · App Development

# Knowledge & Skills

ACA-Applied
K Knowledge · what you know
Use the paid path only for what the product must make live while a user waits, and put everything made ahead of time on the free path.
Sort APIs into types: self-serve (sign up and use), partner-gated (need a partner agreement), and open ones whose data you may store; copying a service that has no public API breaks its rules and can get your account banned.
Follow the set order of steps for slow jobs: upload, store, read, generate, poll, show; large files go straight to storage, and because making the result takes longer than the app will wait, the app starts the job and then checks back later (this checking back is called polling).
Prove that one API call works fully from start to finish before building the product, and write no product code until that connection works.
Build the app so no single file the AI creates holds the whole thing, and give each file a stated job.
Run four roles (CEO, CTO, Marketing, CFO) across four parts of the day, with one role leading each part; everyone works all day and the lead changes with each part.
End each part with a named handover to the next part, because teams run into trouble at the handovers, not inside the parts.
S Skills · what you can do
Research a real, similar company and explain what its website teaches you. point w3d4-s1-p1; evidence w3d4-e3
Take a set role (CEO, CTO, marketing, or CFO) and work as a team to a fixed deadline. point w3d4-s4-p1; evidence w3d4-e6
Draw a diagram showing how the code, the database, the AI, and the APIs (ways one app connects to another company's service) fit together. point w3d4-s3-p1; evidence w3d4-e5
Give a one-sentence pitch for the product: we are X, powered by Y, using Z to turn this into that. point w3d4-s2-p1; evidence w3d4-e4
Publish three working pages: a landing page, a gallery that reads from the database, and a password-protected AI agent. point w3d4-s4-p2; evidence w3d4-e1
T Techniques · named subskills
Comparable-Company Study Study one real competitor's page and name what it teaches you about positioning.
Narrow-to-One-Person Segmentation Narrow a customer segment until you can picture one specific early adopter.
One File, One Purpose No single generated file should carry the whole application.
Prove the Pipe First Get one real API call working end to end before writing product code around it.
The Async Spine Upload, store, read, generate, poll, show — the six-step pipeline for any live generation call.

AI

# AI Concepts

ACA-AI.Concepts
Aligned with AI4K12 · Five Big Ideas in AI View framework →
AI4K12 · Five Big Ideas in AI

Eight Brand Decisions First

Decide the brand before building by making eight brand decisions, then paste the same brand block at the top of every prompt (every instruction given to the AI).

AI4K12 · Five Big Ideas in AI

Defend a Model in Layers

Defend a model in layers rather than one polite request, stacking them in order: an instruction in the system prompt (the AI's starting setup), a guard that blocks a matching reply, a second model that checks the whole conversation, and a list of banned inputs; each layer catches what the one before it missed.

AI4K12 · Five Big Ideas in AI

Insecure Output Handling

Explain insecure output handling (a known AI risk, listed as OWASP LLM02): the app trusts the model's reply and passes it on without checking it, which can leak private information the app holds or run as code if the reply goes straight to a command line or a database.

AI4K12 · Five Big Ideas in AI

Model Denial of Service

Explain model denial of service (a known AI risk, listed as OWASP LLM04): someone floods a model with many demanding requests, which slows it down or takes it offline for everyone else.

AI4K12 · Five Big Ideas in AI

Two-Rail Money Model

Explain the two-rail money model: work made by hand inside the app is paid for by the subscription, and work the app makes live through an API call is charged to the card every time it happens.

# AI Competencies

ACA-AI.Competencies
Aligned with UNESCO · AI Competency Framework for Students View framework →
AI Techniques and Applications
4.3.1

Citizenship in the AI Era

Developing awareness of AI's role in society and one's responsibilities and opportunities as an AI-aware citizen.

# AI Skills

ACA-Tools.Skills
Aligned with Derived from course activity
Derived from course activity

Prompt Engineering for AI Agents

Iterating a prompt/system instruction until an AI agent's responses stop sounding generic and actually reflect the specific product or context. Evidence: "Does this information sound generic? Or does it sound actually tailored to the website?" (app-development, w3d3).

Derived from course activity

Deployment & Secret Management

Setting up a dev environment and deploying live while keeping API keys out of source, held as environment variables. Evidence: "SETTING UP A DEVELOPMENT ENVIRONMENT: 1. Make a folder on your laptop..." and "holding the API key as an environment variable" (api/app-development).

# Competency Goals

ACA-AI.Goals
Aligned with UNESCO · AI Competency Framework for Students View framework →
UNESCO · AI Competency Framework for Students
CG4.1.4.2

Skills to Construct AI System

Understanding the steps, expertise, and resources needed to build an effective AI solution to a problem.

UNESCO · AI Competency Framework for Students
CG4.1.4.3

Assessing Data, Algorithm, and Compute Needs

Evaluating what data, computational methods, and computing power are required for a proposed AI solution.

UNESCO · AI Competency Framework for Students
CG4.3.1.3

Self-Actualisation as AI Citizen

Identifying ways to contribute meaningfully to responsible AI development and deployment in one's community and career.

UNESCO · AI Competency Framework for Students
CG4.2.2.3

Legal Responsibilities and Compliance

Understanding applicable laws, regulations, and ethical standards governing AI development and deployment.

Deliverable

# The Hackathon

Projects.Briefs

Turn an idea into a working website with a database and an AI agent, and get it online before the deadline. A three-page website (a landing page, a gallery backed by a database, and a password-protected AI agent), plus a portfolio of supporting work: the Lean Canvas (a one-page business plan), the published site, the AI agent, the database, a diagram of how it is all built, the cost workings, and the list of problems found and fixed (the defect log).

Transfers to: Splitting a build into named files with one clear purpose each.

See past cohorts ship this in the Showcase →

# Assessed toward these aims

Projects.Rubrics

This deliverable is where students demonstrate

  1. D Chain live APIs into a working product, and cost it
    Look out for
    • Comparable company research: Names a real, similar company and clearly states what its website teaches, and uses this in the team's own plan.
    • Roles and deadline: Each team member works to a set role (CEO, CTO, marketing, or CFO), and the team gets the product online by the deadline.
    • Architecture diagram: The diagram shows how the code, the database, the AI, and the APIs connect, with each part clearly named.
    • One-line pitch: The pitch states, in one sentence, what the product is, what powers it, and the change it makes (we are X, powered by Y, using Z to turn this into that).
    • Three pages shipped: A live landing page, a gallery that reads from the database, and an AI agent behind a password all work during the demonstration.

# Team roles

Projects.Roles
01

UX engineer

Clicks every button, reads the console, and watches someone from another team use the site.

02

Project manager

Writes down every issue in public and ticks it when fixed.

03

AI/API engineer

Attacks the agent and looks for ways to get it doing something it should not.

04

Programmer

Fixes what the other two find.