Overview
Making a hearing aid is surprisingly complicated: a patient visits an audiologist, an ear impression is taken, a 3D model is designed from it, the shell is 3D-printed, electronics are mounted, quality is checked, and the finished device travels back to the clinic to be fitted. Traditionally each of those steps lives somewhere different (spreadsheets, emails, phone calls, separate software) with a lot of manual handoffs between people who often work in different companies and even different countries. This platform runs that entire journey, from order to finished device, in one connected flow. It's not a demo. It's a live system used for daily production, and it describes itself as "the first complete, production-ready digital workflow for hearing-aid manufacturing."
The Challenge
The real difficulty isn't any single step: it's that the whole process spans separate businesses in separate locations who have to coordinate perfectly, over a workflow that's long, branching and adaptive.
- Connecting three groups who normally struggle to coordinate (audiology clinics, remote modelling specialists and manufacturing labs) around one live source of truth.
- Modelling an ~8-stage workflow (Order Placed → Modelling → Printing → Mounting → QA → Fitted → Closed) that auto-routes each order to the right person and adapts, e.g. automatically skipping "ship the impression" when a clinic has its own scanner.
- Serving distinct roles, each seeing only the actions relevant to them, backed by granular per-stage permissions.
- Doing genuinely technical work in the browser (audiograms, 3D device models, print-bed nesting) for users who aren't engineers.
How I Solved It
I was the first engineer on the team, so I owned the foundation: the architecture, the project structure, and the reusable building blocks the rest of the product was assembled from: a shared typography scale and a large component library that kept a very large app consistent and cut delivery time by roughly 30%. The system models the order lifecycle as explicit, tracked stages, so an order's exact position is always known and auditable, and it automatically routes each order to the next role.
It's a Next.js / React / TypeScript web app talking to a GraphQL backend (Apollo), using Three.js to render 3D device models in the browser, charting libraries for audiograms and analytics, and a Tailwind + Radix component system. Guided order creation covers four device families (BTE, RIC, custom in-ear, ear plugs) across a range of specialised plug types, with audiogram entry and per-ear customisation. Labs batch devices of the same colour/printer into single print jobs with barcode/PDF labels and QA cube-measurement checks, and admins configure the whole catalog (styles, materials, colours, manufacturers, receivers, sound tubes, vents, grips and filters) that flows into the order forms.
Key Features
- Guided order creation for 4 device families (BTE, RIC, custom in-ear, ear plugs) plus specialised ear-plug types, with audiogram entry, visual graphs and per-ear customisation.
- Full order traceability: a visual timeline of every order's exact stage, a complete who-did-what action log, and left/right-ear status at a glance.
- Role-based access for 7 roles (Audiologist, Lab Manager, Modeler, Technician, Reviewer, Manager, Patient Care Coordinator) with granular per-stage permission groups.
- 3D & manufacturing: in-browser 3D model viewing (Three.js), AI-assisted modelling with a 3Shape integration, and AI slicing/nesting for the print bed.
- Print-job management: batch devices by colour/printer, barcode/PDF labels, QA verification via cube measurements, and bulk file downloads.
- Company & lab administration: branches, printers, holiday calendars and daily capacity; a configurable product catalog; plus appointments, repairs & warranty flows and a live analytics dashboard.
- Built-in searchable documentation with guided tours; full English/Turkish throughout.
Results & Impact
This is the operational backbone of a real manufacturer: a live, actively-developed production system (a maintained release history, currently around v1.34 in early 2026) rather than a prototype. As the first engineer on the project, I set the architecture, structure and component conventions the whole platform and team kept building on, and the reusable component and typography system I built cut delivery time by roughly 30%. It spans roles across clinics and labs working together in real time, an auto-routing workflow, guided configuration across device families and plug types, and broad coverage of ordering, labs, companies, print jobs, patients, appointments, repairs, analytics and admin: an unusually broad, deeply-connected system.