JOBCOOL

TECHNEST Capstone Project

Weeks 9–12 · Team AI Product

Objective

Over four weeks (9–12), teams of 2–4 students evolve the full-stack Next.js + Neon + Vercel codebase they built individually in Weeks 1–8 into a deployed multi-user AI product that serves a real audience beyond their own classmates.

The capstone is the bridge between "I can ship a personal site" and "I can ship something strangers would use".

Career · What 'multi-user' means here

Your Week-8 site had exactly one user: you. Your capstone must work for strangers, end-to-end. That means: sign-up (not just sign-in), onboarding, data isolation per user, thoughtful empty states, and a reason for a stranger to come back tomorrow.

Team Composition

• Team size: 2–4. Groups of 3 work best.
• Roles are soft. Every member must commit code. No "non-technical" seats.
• Form by Week 9 Day 2. The instructor publishes a sign-up sheet on Day 1 with suggested pairings based on Week-1-to-8 styles.

The Three Tracks

Pick one track; the team may combine elements but the spine should be one theme so the user story is crisp.

🏫 Campus Life Track

Products that serve the daily life of a university community. Low friction to get classmates to try it, which means real user feedback within the 4-week window. Examples:

1. Lecture-note swap — signed-in students upload notes, AI extracts topics + summaries, classmates discover relevant notes by course code.
2. Roommate matcher — personality + schedule profile, AI surfaces compatibility signals, safe opt-in chat.
3. Campus event radar — scrape / ingest event feeds from student societies, AI personalises a weekly digest, Slack / email delivery.
4. Dorm chore + bill splitter — Slack-integrated, AI-nudges for overdue turns, fair-split math for shared utilities.

🌱 Personal Growth Track

Products where the user's state matters — logs, history, streaks. Teaches teams about private data, consent, and longitudinal UX. Examples:

1. AI habit coach — daily check-in, weekly LLM-authored review, pattern callouts ("you skip running when you sleep <6 h").
2. Reading companion — upload PDFs / articles, AI quizzes you on spaced-repetition intervals, tracks true-vs-guess on your notes.
3. Journaling assistant — private journal with optional AI reflection; every Sunday a gentle "here's what you wrote about this month" email.
4. Workout planner — profile-driven weekly plan generator, logs per set, Sunday review of volume progression.

🎨 Creative Tools Track

Generative products with a clear loop: input → AI output → user iterates. Most visually demo-able at the showcase. Examples:

1. AI short-story studio — collaborative writing with AI continuation, branch / merge like Git for narrative choices.
2. Music-prompt playground — structured prompt UI for Suno / ElevenLabs, sharable rooms, voting on outputs.
3. Comic-panel generator — script → panel layout + image generation + dialogue placement → downloadable PDF.
4. Resume-video script maker — turn a Week-8-generated CV into a 60-second narrated video script + optional TTS.

Project Requirements (non-negotiable)

Every capstone, regardless of track, must include:

• Authentication — multiple real human users, not a demo account. Use Neon Auth or equivalent.
• Persistent state per user — Neon Postgres as primary store; per-user data isolation verified.
• At least one AI feature that is meaningfully in the critical path — removing the AI should break the core value. No "AI-washing" where the LLM is an optional side panel.
• Real-time notification — Slack or email, triggered by a user action.
• Deployment to Vercel on a shared team repo. Merge via pull request; no direct-to-main.
• A one-page landing page explaining the product to a stranger in under 30 seconds — hero, problem, demo video, sign-up.
• A working /feedback form that persists to Neon and pings the team's Slack.

Suggested Stack (defaults)

Unless a track strongly demands otherwise, teams ship on the stack learned in Weeks 1–8:

Layer	Default
Framework	Next.js 14 App Router + Tailwind
Hosting	Vercel
Database + ORM	Neon Postgres + Drizzle
Auth	Neon Auth (Google + GitHub providers)
File storage	Vercel Blob
LLM in-product	Gemini 2.5 Flash for cheap calls; Claude for reasoning / agents
Notifications	Slack Incoming Webhooks or Resend for email
PDF / docs	Typst via skill (where applicable)

Teams may deviate only with written instructor approval and a clear reason ("we need vector search → add pgvector"; "we need realtime → add Supabase Realtime").

Timeline

Week 9 — Kickoff, Scope, Repo

• Day 1. Team formation; track selection; 30-min product brief (one page: problem, user, user journey, core AI feature, success metric).
• Day 2. Repo scaffolded; Neon + Vercel provisioned; CI green; 1 commit from every team member landed.
• Day 3. Core data model + auth + one end-to-end vertical slice (sign up → take one action → see result) deployed to production.
• Check-in: instructor reviews each team's vertical slice URL at end of Week 9. Pass / fail — teams that fail redo in Week 10 with a tightened scope.

Week 10 — Build the Critical Path

• Finish the AI-in-the-critical-path feature end-to-end.
• Add the notification trigger.
• First external user test: someone outside the class tries the product; team watches the session (Zoom screen share is fine).
• Log observations in a shared RESEARCH.md.

Week 11 — Harden, Polish, Market

• Landing page + onboarding.
• Feedback form + /bugs channel.
• Accessibility pass: keyboard nav, contrast, error states.
• Five real users sign up by end of Week 11 (not team members).
• Light marketing: a LinkedIn / WeChat / small-group post explaining the product.

Week 12 — Demo Day

• 5-minute live demo per team.
• 3-minute Q&A with instructor + invited guests.
• Submission: live URL + repo + 2-minute recorded video + 1-page post-mortem.

Demo Requirements

At Demo Day every team will present, in this order:

1. Problem statement — 30 seconds. Who hurts? How often?
2. Solution — 30 seconds. The product in one sentence.
3. Tech stack — 30 seconds. What from Weeks 1–8 did you reuse; what's new.
4. Live demo — 3 minutes. Sign up a fresh user on stage; walk through the critical path; land the notification moment.
5. Challenges faced — 1 minute. One technical, one team, one user.
6. What's next — 30 seconds. Honest: is this worth keeping alive past Week 12?

Recording requirements: ≤ 2-minute clean video linked from the landing page and from the team's shared repo.

Evaluation Criteria

Dimension	Weight	What "excellent" looks like
Functionality	30%	Every promised feature works end-to-end on the live site during the demo. No "this part is flaky" disclaimers. Real external users ran the critical path at least 5 times pre-demo.
AI in critical path	20%	The AI feature cannot be removed without the product losing its core value. Thoughtful prompt / system-prompt design, visible in the repo as files (not baked into code).
Code quality + collaboration	15%	Commit history is healthy: every member shipped real features, PR reviews happened, no secrets in git. Naming is consistent; no dead code.
UI / UX	15%	Mobile works. Empty states exist. Error messages are in human language. Onboarding surfaces value in <2 minutes. Accessibility basics pass.
Deployment + reliability	10%	No 500s during demo. Vercel logs are clean. Secrets only in env vars. Database migrations documented and reversible.
Presentation	10%	All team members speak. The problem lands in the first 30 seconds. Demo is rehearsed — no "let me find the button". Post-mortem is candid, not marketing copy.

Instructor Role

• Provide guidance, do not build for them. Answer "how should we approach X?" questions, not "can you write this for us?".
• Unblock early. One 30-minute sync per team per week, on a fixed day, to catch scope creep and infrastructure issues before they rot.
• Review progress weekly. End-of-week async check-in: URL link + two-sentence summary of what shipped.
• Enforce the bar. Teams that don't ship a vertical slice by end of Week 9 must re-scope; teams that haven't ten real users by end of Week 11 get a one-on-one about what to cut.
• Celebrate the finalists. Demo Day is public. Invite the school's career-services team, at least one local tech employer, and alumni from previous cohorts.

Deliverables Checklist

• Live URL (Vercel production domain).
• Public GitHub team repo with README + setup instructions.
• Landing page with problem + solution + demo video.
• 2-minute demo recording (unlisted YouTube / Vimeo / self-hosted fine).
• 1-page post-mortem (what went well, what hurt, what you'd cut).
• List of five external users (with permission to cite first names).
• Five-minute live demo on Demo Day.

A Note from Your Instructor

The cohorts before you each shipped one product that surprised me. A journaling app whose AI I actually still use. A roommate-matching tool that ended up matching two teammates themselves. A comic-panel generator that a local children's museum reached out about. Your capstone is not a class project — it's a real artefact with your real name on it. Aim accordingly.

If nothing else, your capstone should be the first thing you pin to the top of your GitHub profile in January.

Career

Past-cohort capstones have been cited in successful internship applications at Anthropic, Vercel, ByteDance, and three NZ startups. A deployed team-built product that solves a real problem is worth 10 side-projects built alone.

See Curriculum Outline for the overall assessment breakdown across the 12-week programme.