Selected Work

Systems, tools,
and working prototypes.

2026 - Research Audit . Strategy Writing . Web Report

Audit Report

A research-led audit and recommendation piece, reframed for public viewing.

Public-safe note

Shown here in public-safe form, with the research method, evidence handling, and recommendation structure preserved.

Strategy document

Audit to report

Executive summary

Market landscape

Expectations

Audit findings

Gap analysis

Recommendations

Appendix

research docs

08

priority actions

04

appendix tracks

07

Decision view

Executive readout structure

anchored report

Primary narrative

Clear thesis first, supporting evidence second, so the page works for both quick reads and deep dives.

Recommendation block

Sequenced into immediate, next-phase, and longer-term actions instead of a flat list of ideas.

Appendix system

Separated source material, comparative tables, and audit observations into follow-on evidence sections.

Research inputs

merchant reviews

competitor matrix

UX audit notes

source references

Delivery format

executive summaryappendixanchorsscorecardsevidence

Inside the build

What it does

6 modules

01

Market research

Mapped the shipping app landscape, benchmarked merchant expectations, and identified a time-sensitive acquisition gap created by market change.

02

Live product audit

Reviewed the current app experience in a staging environment, documenting onboarding friction, product trust issues, workflow breaks, and support signals.

03

Deep synthesis

Condensed multiple research threads into a single executive narrative, separating fast leadership reads from the deeper evidence appendix.

04

Recommendation matrix

Structured the output into immediate fixes, medium-term UX/platform shifts, and longer-term growth moves so the report could guide sequencing, not just commentary.

05

Web-based delivery

Published the work as a clean report site with anchored sections, appendix navigation, and a reading flow designed for executives and follow-up contributors.

06

Evidence trail

Kept findings traceable through supporting documents, comparison tables, audit notes, and source references rather than presenting unsupported opinions.

System shape

Competitive analysis /Merchant expectation research /Staging UX audit /Executive writing /Information architecture /Decision narrative design /Appendix systems /Static web publishing /Evidence traceability /Recommendation sequencing /Source synthesis /Portfolio-safe reframing

What's shown here

Portfolio copy uses generic framing instead of client-facing internal language.

Sensitive operational context is removed from the showcase description.

The emphasis is on research method, decision support, and report design.

Private repo and live deployment references are intentionally omitted.

2026 - Manifest V3 . Vanilla JS . Chrome Side Panel

Operations Layer

A working operational layer for triage, risk, and queue movement inside a live support environment.

Public-safe note

Shown here as a public-safe reconstruction with generic labels, mock counts, and reworked flows.

Inside the layer

A working model of how triage, guarded actions, and reporting sat inside one continuous operator surface.

Presented as a public-safe reconstruction. Labels, counts, and flows are reworked, but the operating logic stays intact.

Support Feedticket and reply flowQueue Enginepriority scoringPriority Viewranked operator readOps Consolebooking and state dataRisk Gateaction validationAction Surfacegated inline actionsComms Layercall and queue dataState Trackercross-system captureDaily Summarynormalised KPI output

What it does

Queue visibility

Surfaced time-sensitive work across fragmented sources into a single ranked view. Urgency scored by cutoff proximity, reply state, and tracking status - not by operator intuition.

Action gating

Validation layers checked operational preconditions before permitting irreversible dispatch actions. Errors caught at the system layer, not after the fact.

Embedded context

Account, shipment, and address data surfaced directly in the operator's active workspace. No tab-switching during live decisions.

Cross-system reporting

Automated KPI capture across disconnected platforms. Normalised into a structured daily summary without manual compilation or export.

Working logic

01Surface the right signal at the right moment - not everything available
02Gate destructive actions behind state checks, not operator memory alone
03Context should travel with the operator, not live in a separate window

Build notes

Chrome Extension MV3Content Script InjectionCross-origin MessagingBackground Service WorkerPriority Queue AlgorithmCross-system Normalisation

Inside the build

What it does

6 modules

01

Priority board

Ranked active work by urgency, elapsed time, and reply state so teams could act from one live queue instead of bouncing between tabs.

02

Delay intelligence

Calculated business-day ETA drift and surfaced exception states for follow-up work without exposing live carrier or customer details.

03

Queue monitor

Watched live support load with threshold alerts, wait-time emphasis, and team-status visibility in the same panel as operational work.

04

Guarded actions

Wrapped high-risk actions in validation, cutoff checks, confirmation steps, and status gating so common workflows stayed fast but safe.

05

Context tools

Added fast utilities like quote autofill, clean-copy helpers, note cleanup, contextual address routing, and quick navigation shortcuts.

06

Reporting pipeline

Orchestrated hidden-tab capture, cross-system aggregation, and name normalization to generate daily or weekly KPI snapshots from multiple sources.

System shape

Background service worker /Content scripts /Persistent side panel UI /Message passing /DOM scraping /MutationObserver routing /Chrome storage sync/local /TTL caching /Hidden-tab automation /Cross-system aggregation /Feature toggles /Manifest V3 architecture

What's shown here

All labels and metrics shown here are mock values.

Company names, internal tools, and endpoints were removed.

No screenshots from the production extension are used.

The case study focuses on build logic, not client IP.

2025 - Canvas API . Vanilla JS . Sprite Animation

Room for Two

A walkable pixel room built for two - shared space, local/firebase memory. Both players move independently and can interact with objects in the room together.

WASDMove
EInteract
Tristan

Tristan

Juju

Juju