AI in products and development
AI changes both what an electronic product can do and how it is developed. At ProtoMakers, AI already runs inside a client product, and the code we write goes through AI review as a standard part of the development process. We offer both to you.
What we do
- Edge AI: models running on the microcontroller or a small computer inside the product, with no cloud dependency
- Fault detection and predictive maintenance from sensor data
- Real-time signal, audio and image analysis
- Connecting products to large language models (LLMs) through the cloud for a smart user interface
- AI-accelerated firmware development and documentation
- AI-assisted code and design review as an extra layer of checking
- Automated hardware-in-the-loop (HIL) tests, with AI analyzing logs and measurements
When to call us
- You want the product to detect a fault before it happens
- There is a lot of sensor data and nobody has time to analyze it
- You need to shorten development time without compromising quality
- Every new code version breaks something that used to work
What you get
- A feasibility check: does AI really add value to your product
- A model running on the hardware, with accuracy, run-time and memory measurements
- An automated test setup you can run on every version
- An organized test report for every version
FAQ
Is AI really right for my product?
Not always. Sometimes a simple algorithm does the job better and cheaper. That is why we start with a short feasibility check and tell you honestly what came out of it.
Does AI-assisted development mean the code is written by a machine?
No. An engineer designs, writes and checks every line. AI tools speed up the work and add a layer of review, but the responsibility stays with the engineer.
What are HIL tests?
Hardware-in-the-Loop: the real board is connected to a setup that simulates its environment, and automated scenarios run on it. Every version is tested on the hardware, not only in simulation.
Examples from the field
Ultrasonic medical nebulizer
- Challenge
- The piezo element's resonant frequency shifts with heat during treatment. If the controller doesn't follow it, the mist weakens or disappears.
Remote firmware updates for a LoRa field network
- Challenge
- Update the firmware of about 150 units already in the field, over a slow LoRa channel with a 1% duty-cycle limit, without changing a single hardware pin and without bricking devices.
IMU sensor unit at 2 kHz
- Challenge
- Read an IMU sensor and transmit the data to the computer 2000 times a second, with less than 500 microseconds per cycle.
Have a project on the table?
A free 30-minute intro call, no obligation.