Mohamed Sallam
CTO — AI & Embedded Systems
Responsible for the firmware, the inertial sensing and orientation pipeline, and the device's real-time behaviour. Writes the engineering field notes.
About
The Align is a dental handpiece angle-guidance device developed to support dental training and real-time technique guidance. It is designed and built in Egypt by a three-person team, and sold direct.
The Align is a small device that fits onto a dental handpiece and shows its working angle in real time. You choose a deviation limit before you begin. While you work, the device compares the handpiece's current orientation against the position you started from and tells you when you have moved past the limit you set.
It reports through three channels, so you are not forced to look at a screen at the wrong moment: a 128 × 128 colour display, a colour LED, and an audible tone. Everything is driven by a single button.
Holding a consistent handpiece angle is a skill that is difficult to self-assess. The difficulty is not seeing the angle once — it is holding the same angle through a long procedure, and knowing whether today's angle matches yesterday's. Wrist and forearm position drift as the hand tires, and the change is usually too gradual to notice from the inside.
A student learning a preparation can feel that something is different without being able to say what changed, and an instructor watching from across a bench can only correct what they happen to see. The Align exists to make that drift visible at the moment it happens, to the person holding the handpiece.
This matters more than the feature list, and we would rather state it plainly than leave it to be inferred.
The Align is an embedded real-time system. Orientation comes from an inertial measurement unit — accelerometer and gyroscope — read continuously and combined by a sensor-fusion filter, because neither sensor is sufficient alone: the accelerometer is stable over time but noisy under movement, while the gyroscope is smooth in the short term but drifts.
Around that core, the firmware carries the parts that make a sensor usable in a real room rather than on a bench: an offset calibration taken while the device is stationary, a running check on whether each sensor's output is physically plausible, outlier rejection, and long-term drift tracking. Display, sensing and logic run as separate tasks so that redrawing the screen cannot delay a reading.
We are not publishing the filter parameters or the specific processing pipeline here. The engineering reasoning behind the design decisions is written up in our field notes.
We think the distinction between three different kinds of evidence is worth being explicit about, because they are routinely blurred in product marketing.
Testing performed by our own team, on our own builds, using our own methodology. All engineering testing performed on The Align to date falls in this category.
Testing by a third party with no stake in the result. We have not commissioned this, and we do not describe our figures as certified, verified, or laboratory-tested.
A study measuring patient or training outcomes. We make no clinical efficacy claim of any kind.
We do not currently publish a numeric accuracy figure. We would rather publish nothing than publish a number whose methodology we have not yet documented. A field note setting out the measurement protocol and its results is in preparation; the figures will appear there first, with the method attached, so they can be checked rather than trusted.
The Align is the commercial development of a research prototype our team built and published on. The paper is peer-reviewed, open access, and the dataset behind it is publicly available.
Intelligent Dental Handpiece: Real-Time Motion Analysis for Skill Development
Author affiliations span the Arab Academy for Science, Technology and Maritime Transport in Alamein, Heriot-Watt University Dubai, and Ajman University. The research received no external funding and the authors declared no conflicts of interest.
A prototype handpiece captured motion on a dental manikin while practitioners worked. Machine-learning models were then trained to classify that motion into three deviation bands. The dataset comprises 3,720 records from 61 practitioners.
That motion data from a dental handpiece can be reliably sorted into deviation states by lightweight machine-learning models. Classification accuracy across the six evaluated model configurations — logistic regression, random forest, linear SVM, polynomial SVM, RBF SVM and a neural network — ranged from 98.52% to 100%.
This is the part we want to be unambiguous about, because the figure above is easy to misread.
Our own quantitative testing of the shipping device, with its methodology, is still to be published. Until it is, we publish no numeric accuracy figure — see Testing and validation above.
The Align is built by three people. Roles below describe what each person is responsible for on this product.
CTO — AI & Embedded Systems
Responsible for the firmware, the inertial sensing and orientation pipeline, and the device's real-time behaviour. Writes the engineering field notes.
COO
Responsible for operations, production runs and order fulfilment — including the batch process by which each device is built and checked before it ships.
Dental / Clinical Lead
Responsible for clinical relevance and the training workflow, and reviews the dental content published here for accuracy.
The Align is an independent dental angle-guidance product and is not affiliated with Align Technology or Invisalign. Where a longer name is useful — in search results, in citations, or anywhere the context is not already clear — we use The Align Dental Angle Guidance System.
Email hello@thealign.tech. Orders are confirmed by phone before anything is built or shipped, and paid cash on delivery. We are based in Egypt and currently deliver within Egypt.