We are looking for a developer to build a simple and reliable desktop application for a medical practice (orthodontics) that automates the creation of consultation reports from recorded audio. The project will be developed in two phases: Phase 1 is a functional local prototype on a single computer, and Phase 2 will allow scalability with multi-computer usage and file sharing.
Core workflow: record consultation → transcribe audio → transform into structured documents → review & edit → export to Word/PDF.
Main features include: (1) Audio recording with a simple interface (Record/Stop buttons), using an external microphone and local storage. (2) Transcription using Whisper or faster-whisper, with strong French support and preferably local processing. (3) AI text transformation to generate a structured consultation summary, a patient letter (simple language), and a referring doctor letter (medical language). The system must support editable templates and custom prompts, and allow either local LLM or API-based LLM. (4) Review interface with editable fields and clear separation of outputs. (5) Automatic generation of a .docx file based on predefined templates. (6) A clear “Export to PDF” button after validation.
Phase 2 (not required for the prototype but must be anticipated in the architecture): ability to share and access files across multiple computers, using simple solutions such as a shared local folder or lightweight local server. Avoid complex cloud dependencies unless optional.
Technical constraints: desktop application (Windows priority), local-first approach, GDPR-conscious design, modular and scalable architecture.
Suggested tech stack (flexible): Python backend, faster-whisper, LLM (local or API), python-docx or docxtemplater, UI with PyQt or equivalent.
Deliverables for Phase 1: a working local application (single computer), simple UI (Record/Stop, Generate, Edit, Export PDF), template system, installation guide, and full source code.
Important notes: final outputs must be in French, communication with the developer can be in English, and UX simplicity is critical for medical staff usage.
Optional bonuses: speaker separation (doctor vs patient), automatic file naming, and readiness for future integration.
To confirm that you have carefully read the full description, please include the word “bird” translated into the target language of the application in your response.
Final objective: build a tool usable daily in a medical practice, with a clear path toward a multi-user and commercially scalable product.
Apply tot his job
Apply To this Job