A four-language telemedicine platform in 21 days.

Pilgrims at one of the world’s largest annual gatherings; doctors volunteering from other time zones; networks that barely work. In three weeks we built and deployed a platform with patient, doctor, and admin portals in English, Arabic, Urdu, and Farsi — right-to-left down to the prescription PDFs — a live-video consult that degrades gracefully to audio and then to messaging, and an emergency-screening layer that never depends on an AI model. Records are encrypted per organization with a tamper-evident audit chain. The platform was deployed, demonstrated to the client, and tested in all four languages (3,700+ automated tests). The team reports the pilot was deferred on AI cost and on-the-ground logistics; the next season is the target.

What this proves: regulated-adjacent depth, multilingual rigor, and the discipline to write down what we found — including the emergency-screening defect our own testing caught before any patient could.