CP3405 Design Thinking 3 · TR2 2026 · Course Week 9 · Final Sprint

Sprint 9 — Build the Frontend. Cap Off the Project.

vW31 · Due Sun 2 Aug 2026 · 23:59 SGT Final Demo Day · Mon 3 Aug 2026 Build the Frontend — Make It Look Like a Product Scrum · Agile · AI Tools · TR2 2026
Sprint9 · Final
Tag RequiredvW31
DeadlineSun 2 Aug · 23:59 SGT
Demo DayMon 3 Aug
Sprint focusEvery role presents

🏁 Sprint 9, In One Line: Build the Frontend. Cap Off the Project.

"Nine sprints of real work is behind you. Sprint 9 is where you build the thing that finally shows it — a real frontend, a real product, the thing you point to on Demo Day and say: this is what we built."

This is the clearest instruction of the term: this final sprint is about building your frontend, and using it to cap off the whole project. Everything else — the pipeline, the calibration, the automation — is the engine. This week is where you build the dashboard on top of it and make the whole thing look, and feel, like a real product.

Sprint 8 was about hardening what runs underneath. Sprint 9 is about what the room actually sees. A working pipeline nobody can see clearly is a much weaker Demo Day than the same pipeline presented through a frontend built to show it off.

Sprint 9 exists to answer one question for every team: when we open your project on Demo Day, does it look like something real — a product, not a script?

What Demo Day Evidence Looks Like — And Why "It Ran Once" Isn't Enough

Any team can point to a pipeline that produced output this week. That is not what separates a pass from a distinction on 3 August. The question is: what's the proof this has been a real system, not a one-off?

The Single-Run Trap — a Demo Day failure pattern

A team opens Demo Day by running the pipeline live, it produces a prediction, and they treat that as the whole story.

One successful run proves the code works today. It does not prove the system learned anything, survived any failure, or was ever running without someone watching it. That proof lives in your commit history, your tags, and your calibration table — not in a single live execution.

The Demo Day Evidence Test — what every team must be able to show

By Demo Day, your team should be able to answer yes to each of these, with a screen open to prove it:

• Can you show a github-actions[bot] commit from at least one week where no human touched the pipeline?
• Can you show a moment where an API call failed or was rate-limited, and the fallback handled it — not a human patch?
• Can you show the cumulative calibration table from Week 22 through Week 31 — not just this week's number?
• Can every team member explain, in their own words, the part of the system they own?

If the answer to any of these is no, that's what this week is for — not adding something new, but making one of these answers true.

What "ready" actually means

It doesn't mean a scripted performance. It means every person knows exactly which screen, which file, which commit they're pointing to when they say "here's the evidence" for their own part of the system. That clarity is what makes a presentation feel professional rather than improvised.

The Dependency Chain — Now the Thing You're Proving, Not Building

You've run this chain eight times already. This sprint it isn't a new risk to manage — it's the evidence itself. The chain that ran cleanly, sprint after sprint, is your case for a distinction.

R3 Almanac
R4 Macro
R5 Technical
R8 LLM Synthesis
R7 Human Score
R9 Tag vW31

Same chain as every sprint. The only thing that changes this week is what a broken link now costs you — there's no Sprint 10 to fix it in. One missing link this week shows up live, in the room, on Monday.

R2's job this week is the same as always — watch the chain from Monday, not Sunday — plus one new job: confirm every role knows exactly what they're presenting and can speak to it clearly.

What R2 says this week

"This is it — last sprint. Your pipeline is the engine — now let's build the frontend that shows it off. If something's still shaky underneath, fix that first, then put a real face on it. And everyone: say your two minutes out loud to the team at least once before Monday."

What to Submit by Sunday 2 August · 23:59 SGT
1
Final tag vW31
Exact format: lowercase v, capital W, two digits. This is the record that seals the project.
2
Cumulative calibration table, W22–W31
The full nine-sprint accuracy trend, not just this week's delta. R10 owns this — it is the single most differentiating artefact on Demo Day.
3
Final retrospective + Demo Day role assignment confirmed
retrospective_W31.md committed, and every role knows exactly which segment they present Monday.
Standing Scrum Obligations — Still Enforced, Final Sprint

The mid-week check-in and dependency discipline do not relax because it's the last sprint — if anything, they matter more, because there is no next sprint to absorb a slip.

Mon 27
Sprint Planning — final sprint goal. R1 writes the goal, but this sprint the goal is reliability and clarity, not new scope. DoD includes "Demo Day segment prepared" as an explicit item for every role.
Evidence: sprint_goal_W31.md committed before end of Monday.
Wed 29 ⚠️
Mid-week check-in — still mandatory. R2 contacts every role individually. Confirms status and, this sprint, confirms whether each person is clear on what they're presenting Monday.
Evidence: standup_midweek_W31.md committed by Wednesday 23:59 SGT, naming every role.
Thu 30
Agent outputs confirmed committed. Same dependency gate as every sprint — R8 cannot start until R3, R4, R5 have committed.
Evidence: Commit timestamps on agent files.
Sat 1 Aug
Final team sync. Confirm as a team, start to finish, who is presenting what — and that every segment has its evidence ready and linked.
Evidence: No file to commit here — but R2 should be able to describe on Monday when this happened and what changed as a result.
Sun 2 Aug ⚠️
Final tag and retrospective. vW31 cut, retrospective_W31.md committed, cumulative calibration table finalised before midnight.
Evidence: Tag timestamp before 23:59 SGT. This is the last one — no late window.
Sprint 9 Cadence — Final Sprint
WhenWhat happensR2's responsibility
Mon 27 SPRINT PLANNING
Final sprint goal set: reliability and clarity, not new scope.
Facilitate. Confirm every role knows their Demo Day segment. Commit sprint goal today.
Tue–Wed EXECUTION + CHECK-IN
Final reliability fixes only. Roles begin drafting their two-minute segment.
Contact every role individually by Wednesday. Commit standup_midweek_W31.md by Wed 23:59.
Thu 30 SYNTHESIS GATE
Same dependency gate — R8 waits on R3, R4, R5.
Confirm all agent outputs are committed before R8 starts.
Fri–Sat AUTOMATED FETCH + FINAL SYNC
Pipeline runs on schedule. Team confirms Demo Day roles and evidence links are final.
Confirm the fetch ran without intervention. Organise the full run-through.
Sun 2 Aug · 23:59 SPRINT SEALED — FINAL
vW31 tag cut. Cumulative calibration table finalised. Retrospective committed.
Confirm all DoD items checked. Sign off before R9 cuts the final tag.
Mon 3 Aug FINAL DEMO DAY
Every role presents their two-minute segment, live, with evidence.
Open the presentation. Introduce the team's nine-sprint story before handing to each role.
What Changes From Sprint 8
🆕 Build a real frontend. This is your moment to make it look like a product.

Several teams have asked whether they can build a frontend this sprint. The answer is: please do — and don't hold back. Nine sprints of solid data and a working pipeline deserve to be seen as something real, not scrolled past in a terminal window. If you've got the system working, spend real time this week turning it into something that looks like it could ship — a live dashboard, a clean results page, your predictions and calibration history rendered so it actually looks like a product someone built to use, not a script someone ran.

This is the best way to prove everything else you built actually works, because a reviewer believes what they can see clearly far more than what they're told about. Two teams walking in with the same accuracy numbers will not read the same on Demo Day if one of them opens a polished dashboard and the other opens a JSON file.

If you have the bandwidth, this is genuinely worth the rest of this week. One real guardrail: don't let the frontend become an excuse to stop checking the backend still runs clean — a beautiful dashboard sitting on top of a broken pipeline is worse than no dashboard at all. But if your pipeline is solid, go build something you'd actually be proud to open in front of the room.

🆕 New: every role presents their own segment

Every role presents their own two-minute segment on Demo Day. This sprint, actually running that sequence out loud as a team — at least once, before Monday — is expected and will show in how the presentation goes.

🆕 New: the calibration table must be cumulative, W22–W31

Not just this week's delta. The full nine-sprint trend is the artefact that shows whether the system actually learned — it is the hardest thing to fake and the easiest thing to verify live.

Still in effect: mid-week check-in is a graded artefact

standup_midweek_W31.md committed by Wednesday 29 July 23:59 SGT is still mandatory, named per role, with any blockers and R2's response action documented.

Sprint 9 Roles — Final Sprint Checklist by Role
R2 · Scrum Master
Demo Day owner this sprint
Mid-week check-in still mandatory. Confirm every role is clear on their segment and evidence. Run the final team sync. Open Monday's Demo Day presentation.
A segment you know cold goes smoothly Monday. One you're still figuring out live doesn't.
R1 · Product Owner
Final Sprint Goal
Write the sprint goal Monday — reliability and clarity, not new scope. DoD includes "segment prepared" per role.
R3 · Almanac
→ Research Analyst
Final commit by Wednesday, same as every sprint. Know your two-minute segment cold.
R4 · Macro
→ Macro Analyst
Final commit by Wednesday. Prepare your Demo Day segment.
R5 · Technical
→ Data Analyst
Final technical read committed before Wednesday. Prepare your segment.
R6 · Data
→ Data Engineer
Confirm the Friday cron fires one last time without intervention. This is your strongest piece of Demo Day evidence. If your backend is solid, this is your week to build a real frontend for it — make your results look like a product, not a log file. Usually your build to lead.
R7 · Human Score
→ Senior Analyst
Final Wild Card, sourced and specific. Be ready to explain, live, one moment your judgement changed the team's call.
R8 · LLM Operator
→ AI/ML Engineer
Final synthesis run. Be ready to show, live, what happens if one provider fails — the fallback is your Demo Day proof point.
R9 · GitHub Lead
→ DevOps Engineer
Merge all branches Saturday. Cut the final vW31 tag before Sunday midnight. Have the Actions run history open and ready Monday.
R10 · Calibration
→ Product Analyst
Build and finalise the cumulative W22–W31 accuracy table. This is the artefact every team will be judged on hardest — know it cold.
Sprint 9 Timeline
MON 27 JUL — SPRINT PLANNING
Sprint 9 opens · Final sprint goal: reliability + clarity
Sprint goal committed. DoD includes a prepared segment per role.
WED 29 JUL — MID-WEEK CHECK-IN ⚠️
R2 confirms status and Demo Day readiness per role
standup_midweek_W31.md committed by 23:59.
THU 30 JUL — SYNTHESIS GATE
R3 + R4 + R5 committed · R8 starts final synthesis
Final comparison table committed. R10 finalises the cumulative table.
SAT 1 AUG — FINAL TEAM SYNC
Confirm roles, order, and evidence links as a team
Every role confirms what they're presenting and where the evidence lives. Fix any gaps now, not Monday.
SUN 2 AUG · 23:59 SGT — SPRINT SEALED, FINAL
🏁 vW31 cut · Cumulative calibration table finalised · Retrospective committed
Last tag of the term. No late window.
MON 3 AUG — FINAL DEMO DAY
🎓 Every role presents live, with evidence
The last time the team presents together. Nine sprints, on the table.
Discord Submission Template — Sprint 9 · vW31 (Final)

Post in your team channel before Sunday 2 August 23:59 SGT.

TEAM: Team __ / Practical __
SPRINT: 9 · vW31 · FINAL
GITHUB REPO:
RELEASE LINK:

━━━ SM CHECK-IN (R2) ━━━
standup_midweek_W31.md committed by Wed 29 Jul: Y/N
All roles clear on their Demo Day segment and evidence: Y/N
Final team sync completed: Y/N · when:

━━━ SPRINT GOAL ━━━
Our Sprint 9 goal was:
Did we meet it: Y / Partial / N

━━━ PREDICTION ━━━
File committed (prediction_W31.json or .md): Y/N
SPX direction + confidence:
NDX direction + confidence:
IWM direction + confidence:
All 11 sectors covered: Y/N
Human Score Wild Card applied: Y/N
Wild Card reasoning:

━━━ PIPELINE ━━━
Automated fetch ran Sat morning SGT with no human intervention: Y/N
GitHub Actions run URL:
All branches merged before final tag: Y/N

━━━ AGENTS (all committed before R8 synthesis) ━━━
Almanac R3 committed by Thu: Y/N
Macro R4 committed by Thu: Y/N
Technical R5 committed by Thu: Y/N

━━━ AI SYNTHESIS ━━━
LLM 1 via API: Y/N · model:
LLM 2 via API (fallback): Y/N · model:
Fallback demonstrated working: Y/N

━━━ CALIBRATION — CUMULATIVE ━━━
Cumulative table W22–W31 committed: Y/N
Overall direction accuracy across all 9 sprints:
Did accuracy improve over the term: Y/N

━━━ SCRUM HEALTH — FINAL ━━━
retrospective_W31.md committed: Y/N
Every role R1–R10 clear and ready for Mon 3 Aug: Y/N
One thing this team is proudest of across all 9 sprints:
Sprint 9 Standing Reminders
The mid-week check-in is not a Discord message

Same rule as every sprint — R2 must contact each role member individually and document the result in standup_midweek_W31.md, naming each role with a status line. This is the last time it's assessed. Don't let it slip because it's the final week.

Tag format: vW31 — still enforced, no exceptions this time

Lowercase v, capital W, two digits. This is the final release tag of the term — get the format right the first time.

If the backend is still broken — fix that first, then build the face for it

If your pipeline still needs a manual step, or a piece isn't working, close that gap before you sink hours into the frontend — a beautiful dashboard on a broken pipeline is worse than no dashboard at all. But once it's solid, the frontend is exactly where the rest of this week should go.

The data domain is the training environment. Scrum is the lesson.

You were never assessed on whether you predicted the market correctly. You were assessed on whether your team ran a professional delivery cycle — planned, executed, reviewed, and retrospected — with a system that improved sprint over sprint. Sprint 9 and Demo Day are where that whole story gets told at once.