// Viewport below spec
This dossier needs a wider display.
▶ Rotate your phone to landscape
or open on a desktop for the full experience
← RETURN TO BOARD
reading···
CASE LB-2026  //  SEQ-03 EMERGENCE  //  tech lead
SEQ-03
/ 05 case files
SEQ-03 // EMERGENCE · CASE FILE ROLE: TECHNICAL LEAD · 7-PERSON TEAM
// File Index · 05
SEQ-01ACCOMPLICE SEQ-02PUBLICIS SEQ-03EMERGENCE SEQ-04VEED SEQ-05DEFRA
[ SEQ-03 // Case File ]
EMERGENCE
Technical Lead · 7-person team · University of Bristol
ROLE: TECHNICAL LEAD
7-PERSON TEAM · SHIPPED
Subject

EMERGENCE — a speculative project for the "Playable Worlds" brief in Advanced Digital Media at the University of Bristol. The team built playable experiences about AI systems that judge human performance. She was Technical Lead on the seven-person team, owning the technical build.

What She Built

The technical core of the team's AI-interview prototype: a real-time, two-player experience where players pick a role and an AI runs the interview.

Candidate — answers questions posed live by an AI interviewer.

CEO — watches the candidate in real time: typing speed, time spent per question, and the answer text itself, all streamed across as it happens.

She built the backend and the two-client state sync that keeps both screens in lockstep.

Architecture
PLAYER PICKS A ROLE ┌──────────────┐ ┌──────────────┐ │ CANDIDATE │ │ CEO │ │ answers the │ │ watches the │ │ AI live │ │ run live │ └──────┬───────┘ └──────┬───────┘ │ poll ↕ state │ poll ↕ state └───────────┬───────────┘ │ ┌─────────▼──────────┐ │ SYNC BACKEND │ │ in-memory session │ │ no DB · ephemeral│ └─────────┬──────────┘ │ ┌─────────▼──────────┐ │ LLM API │ │ AI interviewer │ └────────────────────┘ SHARED LIVE → typing speed · time-per-answer · answer text
Design Notes
SYNC Polling-based — both clients pull shared session state STATE Held in memory; no database LIFESPAN Ephemeral — a run leaves no trace, cleared on completion AI LLM API as the live interviewer

The no-database, ephemeral design wasn't a shortcut — it fit the theme: a performance watched intensely in the moment, then gone.

Where the Team Landed

After testing the interview direction, the team pivoted to a second concept for its final shipped piece — Omni-Triage, an ER scenario on the same theme, where an AI reads a doctor's hesitation as failure. That final build was front-ended by a teammate; her contribution was the technical lead role and the synced interview prototype above.

What It Proves

She can stand up a working real-time, multi-client system with an LLM in the loop — and lead the technical side of a team from concept to a running prototype. The kind of "make the AI thing actually run across two screens" problem, owned end to end.

continue readingnext file → SEQ-04 // VEED
Case LB-2026 · Subject file SEQ-03 · File complete · ←/→ browse