Doctify
Co-founder. An AI clinic product. I owned the architecture and rebuilt the entire application from start to finish.
Doctify is a product for modern doctors. It lets them transcribe consultations, manage the clinic, and move repetitive documentation to AI so they can focus on care.
This was not a clinic product with one AI feature. It has several:
- Live transcription of medical speech, including French, Arabic, and Darija, using Gemini
- Clinical AI chat that can use patient context
- Reports, prescriptions, summaries, and follow up suggestions
- Search over uploaded patient and medical documents
- Patients, appointments, tasks, invoices, team roles, and clinic operations
What I did
I am a co-founder. I was responsible for the architecture and the complete technical rebuild from start to finish. I was also responsible for how the product should behave, not just for implementing a list of features.
The old system was a legacy Laravel product. We rebuilt it as a NestJS, React, and PostgreSQL platform. Doctor, secretary, clinic owner, and platform admin each have their own workflow.
I do this work best when I am given a problem and the authority to create the solution. I am a poor fit for being told to build one exact thing when I can already see that it will not work.
Why we rebuilt
This is the same pattern as Zèya. When a codebase holds several conflicting versions of one idea, and the real source of truth lives in people's heads, AI becomes useless. Patching that kind of system is slower than starting clean. A rebuild is cheap when AI is part of the delivery, and it gives both people and agents one place to work from.
The product later accumulated too many features. That is part of the story too. A rebuild can go too far if nobody is deciding what the product is for.