Planning & Development Workflow
Requirements, Excalidraw system mapping, Variant UI planning, local tooling, Git/GitHub and responsible AI-assisted implementation.
An AI-assisted web engineering competition. Every team receives the same application brief on competition day, then plans it, designs it, builds it, deploys it and presents it — and adapts when the requirements change mid-competition. Three preparation workshop days run first, walking through the exact workflow the competition tests.
| Category | What is assessed | Weight |
|---|---|---|
| Functional Completeness | Does the application meet the provided requirements and function correctly? | 20% |
| Technical Implementation | How effectively were authentication, persistent data, application logic, GitHub and deployment implemented? | 20% |
| Problem Solving & Adaptability | How effectively did the team debug problems and respond to the unexpected requirement change? | 20% |
| UI/UX & Design | Is the application intuitive, coherent, responsive and easy to use? | 15% |
| System Understanding & Planning | Can the team explain its Excalidraw diagram and how the major parts of the system interact? | 10% |
| Production Readiness & Performance | Deployed correctly, connected to a domain, prepared for search discovery, and evaluated with PageSpeed Insights. | 10% |
| Final Demonstration | Can the team clearly demonstrate and explain its solution? | 5% |
AI may generate part or all of your application. There is no requirement to write a specific amount of code by hand. What the rules protect is fairness, not manual effort — so the AI tier limit is the one hard line: free and standard paid consumer plans are allowed, highest-tier and advanced premium access is not. ChatGPT Plus qualifies; ChatGPT Pro does not, and the same principle applies to every other provider.
Coding, debugging, design, architecture and problem-solving help may come from AI or from your own registered teammates — nobody else. Competing teams may not exchange code, prompts, solutions, implementation details or strategies. Public libraries, frameworks and documentation are normal development resources and are never counted as outside help.
No team may start building a solution for the competition challenge before it is officially released, and no pre-built or copied application may be submitted as the team's work. A requirement change arrives mid-competition and must be satisfied — ignoring it is a rule breach, not a design choice. The version standing at the deadline is the version that gets judged.
The workshops and the competition run on one deliberate stack, so nobody spends competition day fighting an unfamiliar tool.
Turning a problem into a structured development plan before writing code: understanding requirements and constraints, breaking a project into features, planning the application structure, splitting frontend and backend responsibilities, and using AI strategically at the planning stage.
Turning the plan into a working application and learning to diagnose problems when it breaks: connecting frontend and backend logic, working with application state and data, implementing and consuming APIs, reading errors, and using AI to investigate rather than blindly rewrite.
The final stage: take the application from localhost to a real production URL, point a domain at it through Cloudflare, submit it to search engines, and measure it with PageSpeed Insights. Deploy, domain, index, measure.
No. The jam is built for participants who are still learning — the workshops introduce the whole workflow first. What the competition asks is that your team can investigate a problem, direct its tools at a solution, test the result, and verify the application actually works. You do not need a team before registering; the team formation page will place you in one.
The working library behind JAM.26: participant-facing lessons and checklists, instructor planning records, and reusable templates for future workshops and competitions. Draft planning files are labeled separately from published learning material so participants can tell what is ready to use.
Requirements, Excalidraw system mapping, Variant UI planning, local tooling, Git/GitHub and responsible AI-assisted implementation.
Firebase Authentication, Firestore persistence, user-owned data, security rules, browser evidence and systematic debugging.
Vercel deployment, Cloudflare DNS, production verification, search discovery, PageSpeed analysis and production debugging.
These filled planning records contain objectives, prerequisites, preparation tasks, lesson timing, demonstrations, exercises, prompt examples, troubleshooting guidance and post-workshop review fields. “TBD” and “Not Started” values remain part of the source planning documents.
Blank structures for future ACM projects. These are working templates rather than event announcements; placeholder fields must be completed and reviewed before publication.
A second copy of the JAM planning files was uploaded under the CTF 3.0 workshop directory. File hashes and content are identical, including the AI web-development curriculum, so those copies are retained as source material but are not published or described as CTF training.
SYS.MSG: CHALLENGE LOCKED UNTIL COMPETITION DAY
Workshops 15–17 September 2026. Competition 19 September 2026. Register your team, learn the workflow, then build without the tutorial.
Register a Team