Engineering Capability
On-Device
Machine Learning
Voice anti-spoofing models trained, quantized to INT8, and deployed to ultra-low-power neural processors. Audio inference at the microphone, never in the cloud.
LCNNINT8 QuantizationONNX → TFLite → NPUASVspoof 2019
Training pipeline
PyTorch-based training pipeline on ASVspoof 2019 Logical Access. Reproducible runs with pinned hyperparameters, dataset versioning, and seed control.
- PyTorch + ASVspoof 2019 LA
- Reproducible training under seed control
- EER metric tracked across epochs
Quantization for ultra-low-power
Post-training quantization to INT8 with a representative calibration set. Model size and inference latency are both cut substantially, with quantization-induced accuracy loss kept to a minimum through calibration.
- PTQ with representative calibration set
- Quantization-induced accuracy loss kept to a minimum
- Memory footprint suitable for embedded NPU
NPU deployment
ONNX export → TFLite intermediate → NPU vendor format conversion. Validated on EU silicon vendor embedded NPU and other low-power accelerators. Inference latency <30 ms per audio frame.
- ONNX as exchange format
- TFLite as intermediate
- Vendor-specific NPU runtime integration