I maintain Matchmaking_site, a JavaScript repository on my @ud-ai GitHub account, github profile. Its README calls the application TDC Matchmaker Dashboard, an internal MVP for The Date Crew (TDC) matchmaking team. It is intended to consolidate client profiles, calculate compatibility matches, compose personalized AI introduction pitches, and maintain log histories.
My Matchmaking_site repository started 2026-06, last updated 2026-06.
The dashboard uses the Guided Connection portal's brand identity, with an editorial Serif-Sans visual language, deep rose and slate-navy color systems, rounded bento-style grids, and glassmorphic micro-animations.
My TDC Matchmaker Dashboard's seeded roster contains 210 detailed profiles: 105 male, 105 female.
The dossier offers Over 30+ structured biodata fields organized into Personal, Location, Work/Education, Identity/Family, Lifestyle/Preferences tabs.
The client database provides a Paginated roster; search by client name or city; filters by gender, status, and city; statistics for total roster, active seekers, onboarding accounts, and matched couples.
My dossier workflow supports Statuses Active, Paused, Matched, Onboarding; status changes update globally; Quick Notes records phone-call summaries.
Notes and status modifications persist across browser refreshes via localStorage sync; local JSON data is merged with browser persistence for notes, tags, and sent proposals.
My male matching logic Filters for female pool; scores age gap (younger), height (shorter), income bracket, kids preferences, religion/caste preference, and marital history.
My female matching logic Filters for male pool; scores lifestyle alignment (diet/drinking/smoking sync), education alignment, income parity, relocation compatibility, family type (nuclear vs. joint), horoscope matching, and partner age parameters.
The matching engine automatically ranks candidates, with score < 45 filtered out by default.
The simulated AI default; local templates generate 3-line client summaries, 2-sentence match reasons, and personalized email pitches.
Live generation supports Google Gemini 1.5 Flash, OpenAI GPT-4o mini, Anthropic Claude 3.5 Haiku; live generation requires developer key in AI Settings.
Match suggestions flag potential value mismatches including dietary differences, smoking versus non-smoking, and pet preferences.
In the match composer, Send Match opens a composition modal with connection header and AI-drafted email; subject and body can be edited; sending shows success toast and documents dispatch.
Drawer logs review outbox emails proposed during the session.
The internal dashboard login credentials are Username matchmaker@thedatecrew.com; password TDC@2026.
My dashboard tech stack is React Vite single-page application; Lucide React; custom Vanilla CSS variables in src/index.css; local JSON profiles at src/data/profiles.json.
npm install
python scripts/generate_profiles.py
npm run dev
npm run build
The development server usually runs at http://localhost:5173.
The README's documentation references say docs/architecture.md covers files, folder tree, and Mermaid data flows; docs/assignment_writeup.md covers technology choices, matching algorithms, AI providers, and assumptions.
The README links to those documentation files using local file:///c:/Users/manup/OneDrive/Documents/Matchmaking_site/... paths; their contents are not included here.