Shakeel Pakwan
A bilingual (English/Urdu) catering and events website for a Karachi caterer, built to turn a menu scroll into a WhatsApp quote request with no framework, no backend, and no dropped leads.
The Problem
Shakeel Pakwan caters weddings, corporate events, and family gatherings in Karachi, and most of it is booked over a phone call or a WhatsApp voice note, not email. The site needed to hold a genuinely large menu (six sections, 80+ dishes) in a way that reads naturally in Urdu as well as English, run fast on a mid-range phone over a mobile connection, and turn a browsing visitor into a quote request without a backend to maintain.
The Starting Point
No existing website. Built from scratch for an already-operating catering business (40+ years in business, per their own site).
The Solution
Seven static pages, Home, About, Catering Services, Menu, Gallery, Contact, and Request a Quote, shipped with zero framework and no build step, deployed straight to Vercel. Every dish on the menu carries its Urdu name set in proper right-to-left type next to the English, across all six menu sections. Contact and quote forms validate in the browser, then hand off a pre-filled wa.me WhatsApp message instead of posting to an email server or CMS. The photo and video gallery is built on the native <dialog> element for a free focus trap and Escape-to-close, plus a scroll-snap slider.
Architecture
Engineering Work
A performance pass cut media weight without touching visible quality. Gallery videos were shipping at 576px wide into tiles displayed at roughly 260px, nearly five times the pixels the layout ever showed, so they were re-encoded to match the real display size and given poster frames, cutting video weight from 8.2MB to 3.1MB (a 62% reduction). Every photo moved from JPEG to WebP, taking site-wide photo weight from 5.7MB to 3.3MB (down 45%) and homepage image weight from 2.59MB to 1.59MB (down 38%) on the first page most visitors land on. Accessibility got a skip link, aria-current on the active nav item, and an accessible mobile menu toggle throughout. Structured data (FoodEstablishment and Menu JSON-LD), per-page Open Graph tags, a sitemap, and robots.txt were added for search.
AI Contribution
No AI is used in this product. It's included in the portfolio for its production-engineering rigor, not for AI usage.
Results
Seven pages, 80+ menu items across six courses in two languages (English and Urdu), and a 38 to 62 percent cut in media weight across homepage images, site-wide photos, and gallery videos, with zero visible quality loss and zero framework dependencies.
Final Product
