pinned
Running
Milk NIR Simulator
🥛
Raw-milk fat, protein and lactose from NIR, in your browser
AI & ML interests
None defined yet.
Recent Activity
View all activity
Organization Card
Amazeng
High-precision data acquisition and Edge AI for industrial automation and smart agriculture.
We are an engineering company from Ankara, Türkiye. We design the full chain between a sensor and a decision: 24-bit data acquisition hardware, embedded Linux HMI panels, IoT/cloud integration and AI models that run at the edge, next to the machine.
What we publish here
Open, small, deployable models for sensor and spectral data, each with:
- ONNX weights with preprocessing built in: raw sensor data in, engineering units out
- Honest model cards: temporal or device-level test splits, limitations stated, nothing hidden
- Attributed open data: every model names its training dataset and licence
| Model | Task | Status |
|---|---|---|
| milk-nir-composition | Raw-milk fat, protein and lactose from NIR transmittance (960–1690 nm) | Released |
| NIR plastic classification | Polymer type from NIR spectra for recycling lines | In progress |
| Load-cell fault detection | Sensor and wiring faults in 24-bit, high-rate weighing signals | In progress |
Where our models run
- Spectral analysis: NIR for milk quality and plastic sorting, multispectral imagery for agriculture
- Industrial signals: load cells, LVDTs and strain gauges in materials testing and process control
- Edge hardware: NVIDIA GPUs and embedded Linux panels, close to the sensor, without a cloud round-trip
Products
ZMA data acquisition (24-bit, 1 kHz) · GDT digital transmitter · Qt/QML industrial HMI panels · materials test machine controllers · milk tank controllers · cloud and IoT data collection
Contact
Working on spectral or industrial sensor AI? We are happy to collaborate on data, calibration and edge deployment.