import { Pipeline } from '@gargantua/core';
import { SchemaRegistry } from './registry';

const pipeline = new Pipeline({
  source: 'enterprise-lake',
  transforms: [
    normalize({ encoding: 'utf-8' }),
    deduplicate({ key: 'entity_id' }),
    enrich({ provider: 'knowledge-graph' }),
  ],
});

async function ingest(stream) {
  const schema = await SchemaRegistry
    .resolve(stream.metadata);
  return pipeline.run(stream, { schema });
}

from transformers import AutoModel
from gargantua.cognitive import Agent

class CognitiveEngine:
    def __init__(self, config):
        self.model = AutoModel.from_pretrained(
            config.base_model,
            quantization='int8',
        )
        self.agent = Agent(
            reasoning='chain-of-thought',
            tools=config.tool_registry,
        )

    async def inference(self, prompt):
        context = await self.agent.plan(prompt)
        embeddings = self.model.encode(context)
        return self.agent.execute(
            embeddings,
            temperature=0.7,
            max_tokens=4096,
        )

model = Sequential([
  layers.Dense(512, activation='relu'),
  layers.Dropout(0.3),
  layers.Dense(256, activation='relu'),
  layers.Dense(num_classes, activation='softmax'),
])

model.compile(
  optimizer=Adam(lr=3e-4),
  loss='categorical_crossentropy',
  metrics=['accuracy', 'f1_score'],
)
Gargantua Group Logo
AI Mastery Program · 9–11 Week Track

AI Mastery
Program

Nine to eleven weeks. One real AI product. A live audience.

For students who want to make, not just consume: identify a real problem, build a working AI product, and pitch it at a public Demo Day.

9–11 Weeks · 1 School Term
36–44 Hours · 2× / Week
Ages 12–16
No prereqs

Developed in Partnership With

Gargantua Education Quest X XPRIZE

Backed by STEMNovation, the original program from Valley Christian Schools that Danny and Hannah Kim started.

Program Overview

The AI Mastery Program is our flagship after-school track. It fits a 9–11 week school term, meets twice a week, and culminates in a public Demo Day. Every team presents their live AI product to a public audience and a panel of guest industry AI and business experts for rigorous, professional feedback.

Students don't learn a tool — they learn how to direct an AI product. Product discovery, context engineering, agent design, AI-assisted development, evaluation, iteration, and launch. The framework stays stable; the specific tools, models, and case studies are updated each term so students are never learning yesterday's skills.

Discover

First third · Find the right problem

Build

Middle third · Turn it into a product

Ship

Final third · Test, refine, and present

Built for an
Industry That
Moves

A curriculum that locks itself to today's tools will be teaching yesterday's skills by the time a student reaches college. The AI Mastery Program is designed around that reality.

The Stable Framework

The spine of the program. Skills, habits, and mental models that stay valuable regardless of which tools win.

  • • Problem discovery & user research
  • • Product specs & context engineering
  • • Agent loops, tools & permissions
  • • AI-assisted product development
  • • Evaluation datasets, rubrics & guardrails
  • • User testing, iteration & launch
  • • Business model, storytelling & pitching

The Evolving Layer

Updated every term by our R&D team based on what's shipping in industry.

  • • AI collaborators (ChatGPT, Claude, Gemini, or next)
  • • Coding agents (Codex, Claude Code, Cursor, Replit Agent, or next)
  • • AI app builders (Lovable, v0, or next)
  • • Product design (Figma and AI-assisted design tools)
  • • Full-stack launch (GitHub, Supabase, Vercel, or next)
  • • Agent frameworks (OpenAI Agents SDK, Claude Agent SDK, or next)
  • • Tracing & evaluation tooling
  • • Fresh industry case studies & guest experts

What Students Walk Away With

Three outcome value propositions, baked into every phase of the program.

The 10× Problem Solver

A student who can identify a problem, mobilize AI to build a solution, and ship something that works will outperform peers by an order of magnitude in whatever they choose next.

Entrepreneurship Fuel

Students ship a real product, test it with real users, build a real business case, and pitch to a real audience. They experience the full arc of taking an idea to a shipped product.

Admissions Differentiation

A shipped product, a demo video, a GitHub repository, a landing page, and a story about what they built, why, and how. The kind of material that anchors a strong application essay.

The Ultimate Pathway

Feeder Program to the XPRIZE Team

This program is the exclusive proving ground for our elite 6-Month AI Incubator. We are aggressively scouting the next generation of visionary tech founders. Only the absolute top-performing AI masters from this program will receive an invitation to join our long-term competitive teams, gaining access to partner resources and the chance to compete on the global XPRIZE stage.

Cadence

Base Program · 2× per week

Two 2-hour sessions per week, after school
9–11 weeks, aligned to the school calendar
36–44 total instructional hours
Culminates in a public Demo Day
Teams of 4–5 · Cohorts of 16–25 (4–5 teams)
No technical prerequisites — portfolio- and mindset-based admissions

Expansion Program · 3× per week

The same 9–11 week arc, with 54–66 total hours. Adds a second user-testing cycle with external users, richer product scope, and more pitch rehearsals. Recommended for schools with existing strong CS or entrepreneurship programs, or cohorts aiming at XPRIZE pathway opportunities.

A Typical 2-Hour Session

0:00–0:10
Standup
What did you work on since last session? What's blocking you?
0:10–0:40
Lesson / Facilitated Activity
New concept, framework, or technique. One idea, well-taught.
0:40–1:30
Build Block
Teams work on their product. Facilitator circulates — unblocks on request.
1:30–1:50
Share-out
One team demos progress. Peer feedback: I-like / I-wish / what-if.
1:50–2:00
Close
Commitments for next session. Quick retro.

The Three-Phase Product Arc

Built to fit a 9–11 week term. Three phases. One team. One shipped product.

Phase 1 · Discover

First third: Find a business problem worth solving, design its agent

Students find a painful problem, write a product spec, and design the agent that could solve it: its context, tools, permissions, human checkpoints, and definition of success.

CORE FOCUS

AI Product Foundations

Team roles, product opportunities, AI capabilities and limits.

Ships: Founder problem log

CORE FOCUS

Context Engineering & Product Spec

User stories, acceptance criteria, structured inputs and versioned context.

Ships: Product spec + context pack

CORE FOCUS

Agent Blueprint

Agent loop: observe, decide, use tools, check results, stop. Add permissions and human approval.

Ships: Agent blueprint + risk checklist

PHASE MILESTONE

Customer Discovery & Lock-In

Live interviews, insight synthesis, market alternatives and team commitment.

Ships: Validated opportunity brief

Phase 2 · Build

Middle third: Build a real AI product and prove it works

Students turn their blueprint into a launchable MVP: prototype the experience, direct coding agents, build an agent loop, then measure it with a real evaluation framework before it reaches users.

CORE FOCUS

Product System & Prototype

Business model, user flows, Figma and AI-assisted prototyping.

Ships: Product flow + clickable prototype

CORE FOCUS

The AI Development Loop

Plan, build, read, run, test, debug: direct coding agents and keep control with GitHub.

Ships: First feature + team repository

CORE FOCUS

Agent Loop, Tools & Evals

Tools, state, structured outputs, guardrails, test cases and a quality scorecard.

Ships: Agent loop + evaluation set v1

PHASE MILESTONE

Launchable MVP

Connect interface, agent and backend; publish a secure, shareable product for testing.

Ships: Live MVP

Phase 3 · Ship

Final third: Evaluate, improve, and launch with proof

Students test their product with real people and adversarial cases, improve the agent loop with evidence, then launch with a founder story, landing page, and public Demo Day pitch.

CORE FOCUS

Real-World Testing & Evals

User observation, edge cases, red-team prompts, rubrics and success metrics.

Ships: User + agent evaluation report

CORE FOCUS

Improve the Loop

Trace failures, improve context and tools, retest, then prioritize what creates real user value.

Ships: Iteration log + launch candidate

FINAL MILESTONE

Launch & Demo Day 🚀

Landing page, demo video, founder story, launch metrics and public showcase with an expert panel.

Ships: Live launch + founder portfolio

Demo Day

The capstone. From the first session, every student is building toward a public, school-wide showcase. Each team delivers a 7-minute founder pitch — hook, problem, solution, live demo, business case, team — to faculty, families, and a panel of industry leaders, including Nick Kim , CEO of Gargantua Group, alongside other AI and business professionals. Students receive the rigorous questions and feedback real founders face when bringing a product to market.

7-minute pitch with live demo
Expert panel Q&A on stage
Written feedback summary within 1 week
XPRIZE pathway invites for top teams

Assessment

No Written Exams

Assessment is continuous, project-based, and performance-oriented. Students are evaluated on what they ship, not what they memorize. A formal feedback report is issued at the end of the trimester — designed to be useful in future applications.

Daily standups Every session
Context & Agent Playbook Early in the first third
Validated Opportunity Brief End of the first third
Product flow + clickable prototype Early in the middle third
Live Agent MVP End of the middle third
User + Agent Evaluation Report Early in the final third
Demo Day Presentation Demo Day
Individual Portfolio By Demo Day
R&D Commitment

A school adopting this is not adopting a fixed curriculum.

They are adopting an ongoing relationship with an R&D team that keeps the program current. The core team operates at the frontier of the AI industry — at Google, Coupang, DeepMind, XPRIZE — and rewrites modules within the trimester when a new model family changes what's possible.

End of every trimester
Tool selections, prompt library, case studies, guest topics.
Annually
Full module content review, sample projects, rubrics, instructor training.
As-needed
Targeted rewrites mid-trimester when a major industry event demands it.

What Comes
Next

The AI Mastery Program is the foundation course. Students who complete it and want to continue can progress into advanced follow-on courses in the Singularity AI Labs sequence — more ambitious products, deeper technical work, external user testing, and pathways into XPRIZE global youth competitions.

Schools can adopt this program on its own, or as the entry point into the full sequence as it rolls out. There is no commitment required beyond this initial track.

Ready to build something real this term?

Ask About the Next Cohort