AACE AI SW Student Home
👤ParticipantACE AI SW Training
ACE AI SW Startup Training

AI Real Project
10-Day Complete Course

It doesn't stop at learning AI tools. In this hands-on course you analyze how a service grows, write design documents, build a real MVP, and present a POC and your final project.

Program Outcomes

What You Should Be Able to Do After 10 Days

AI does not build things for you. You learn to set goals and standards yourself and use AI as a partner for analysis, design, building and testing.

Service Analysis

Reverse-engineer user value, core flows and growth structure with evidence and metrics.

AI-Driven Design

Turn requirements, data, APIs, screens and done criteria into consistent documents.

MVP Build

Use Codex and Antigravity to build a product whose core user flow works.

Testing & Presentation

Use test evidence and feedback, then demo the product's problem, solution and value.

01

Week 1 · Foundation & Guided Practice

AI Development Basics & Core Labs

In week 1 you set up a stable dev environment, then work through service analysis, AI design, document-driven building and backend integration in order.

Day 01Monday

Course Intro & AI Dev Environment Setup

ORIENTATION

Understand the goals and outputs of the 10-day course, and make sure every student starts on the same dev environment.

Morning · UnderstandReview the full schedule, the goals of ACE AI SW Startup Training, and team project and presentation standards.
Afternoon · Set UpInstall and log in to Codex, Antigravity, Claude, Git and browser dev tools, and check they work.
Wrap-up · CheckCreate, edit and run a small file with each tool, and fix any problem setups together.
Today's OutputDev environment checklist, login/run check for each tool, personal learning goal
Day 02Tuesday

Basics & Reverse-Engineering Service Growth

REAL PROJECT 1

Analyze a finished service not as a feature list, but as a structure linking user value, behavior flow and growth metrics.

Morning · BasicsLearn AI collaboration principles, good context, separating evidence from assumptions, and how to check results.
Afternoon · Reverse-EngineerObserve the target service's users, key actions, conversion path, retention and revenue model.
Wrap-up · ShareCompare results across teams and choose the key hypothesis to test in your project.
Today's OutputService benchmarking map, user value flow, key hypothesis and test questions
Day 03Wednesday

AI-Driven Design & gstack

DESIGN

Don't jump straight into coding. Design requirements, data, APIs, screens and test criteria with AI first.

Morning · DesignStructure the problem, user scenarios, feature priorities and failure conditions in documents.
Afternoon · gstackUse planning, engineering and design review flows to find gaps and conflicts in the design.
Wrap-up · ContractConfirm the document contract and done criteria AI must follow when building.
Today's OutputProject requirements, user flows, data/API draft, build done criteria
Day 04Thursday

Antigravity & Document-Driven Building

BUILD

Use design documents as input for code generation, and build in small testable features instead of everything at once.

Morning · PrepareSet the project structure, run commands, build order and the scope of the first vertical feature.
Afternoon · BuildBuild screens and logic with Antigravity and Codex, checking changes in small steps.
Wrap-up · ReviewReview gaps between design and code, run errors, accessibility and exception flows.
Today's OutputA first working feature, build log, design–code trace table and fix list
Day 05Friday

Backend Structure & Apps Script Integration

BACKEND

Separate data responsibilities behind the screen and build a simple server integration with Apps Script and Google Sheets.

Morning · StructureUnderstand responsibilities per layer, data validation, error handling and API request/response contracts.
Afternoon · IntegrateConnect an Apps Script web app to a Sheet and build input, lookup and logging features.
Wrap-up · TestTest normal requests, bad input, duplicate requests and network failures.
Today's OutputData structure, Apps Script API, connected screen, error and duplicate test results
02

Week 2 · Team Project & Delivery

Real Project: Plan · Build · Test · Present

In week 2 your team shares one product goal, leaves a result you can check each day, and finishes a project ready for the final presentation.

Day 06Monday

Real Development Process & Project Kickoff

PROJECT KICKOFF

Agree on your team's problem, users, scope and roles, and make a realistic 5-day plan.

Morning · GoalDefine the project problem, target users, success conditions and the final demo scene.
Afternoon · PlanSet roles, priorities, task dependencies, daily checkpoints and done criteria.
Wrap-up · KickoffPresent your team plan, cut excess scope and confirm the core user flow.
Today's OutputProject Brief, role chart, MVP scope, task backlog and done criteria
Day 07Tuesday

AI-Driven Project Design & Mid-Term Presentation

PLAN & REVIEW

Agree on the screen, data and API structure of the core user flow, and get feedback from peers and instructors before building.

Morning · Detailed DesignDetail screen flows, data structure, API contracts and normal/failure scenarios.
Afternoon · PrototypeQuickly build key screens and interactions and check the core flow.
Wrap-up · PresentPresent the problem, users, design and MVP scope, and sort feedback by priority.
Today's OutputDesign package, prototype, mid-term slides, feedback plan
Day 08Wednesday

MVP Build & Integration Check

MVP BUILD

Build the most important user flow so it works from start to finish, and verify screens, logic and data connections.

Morning · BuildSplit core features across the team and build and integrate them in small working pieces.
Afternoon · IntegrateConnect user input, state changes, data saving and error display into one flow.
Wrap-up · CheckShare progress, blockers and remaining risks, and confirm tomorrow's POC scope.
Today's OutputIntegrated MVP, how to run it, key test results, blockers and fix plan
Day 09Thursday

POC Demo & Feedback-Based Improvement

POC REVIEW

Demo a real working POC, and list the defects and messages you must fix before the final presentation.

Morning · VerifyCheck core scenarios, error flows, data accuracy and a reproducible demo setup.
Afternoon · POC DemoPresent in the order problem → solution → demo → tech decisions, and collect specific feedback.
Wrap-up · ImproveSort feedback into blocking defects, quality improvements and talk improvements, and choose final tasks.
Today's OutputPOC demo, feedback notes, final fix list, presentation storyline
Day 10Friday

Final Project Presentation & Graduation

FINAL & GRADUATION

Clearly show the value of your finished project and how you built it, and define the next steps to grow it.

Morning · Final PrepDo a final check of the demo setup, backup video, slides, run docs and project repository.
Afternoon · Final PresentationPresent the problem, users, solution, live demo, tech decisions, test results and next steps.
Wrap-up · GraduationShare peer feedback and reflections, and organize your project output and learning records.
Final OutputFinished project, run and design docs, test evidence, final slides and retrospective

Core Toolchain

AI Development Tools Used in the Course

The goal isn't to memorize tool names — it's to use each tool safely at the right stage of development.

C

Codex

A development partner that understands the codebase, builds, investigates errors, tests and reviews.

A

Antigravity

A development environment for quickly building and iterating on screens and features from design documents.

AI

Claude

Reviews long documents and alternatives, and structures requirements, designs and presentation messages.

G

gstack

Connects planning, engineering, design and QA reviews into the development flow.

Completion Checklist

Course Completion Criteria

The size of your final output matters less than whether problem, design, build and testing are connected — and whether your team can explain its decisions and reasons.

The core user flow really runs and can be demoed.
You can explain how requirements, design and build results connect.
You have evidence of testing the normal flow and key failure conditions.
You recorded what AI suggested separately from what your team finally decided.
You submit the finished project, docs, slides and a next improvement plan.