Uploaded April 2026 | Updated September 2026, 10 minutes ago

---

Resources:
~~ Understanding ONNX & ONNXRuntime: youtube.com/watch?v=Fxb9QSjrqsI
~~ ONNXRuntime C++ Example: github.com/Vinayak-D/libtorch_onnx_examples/tree/main/cpp_onnx
~~ PortAudio & Sounddevice: python-sounddevice.readthedocs.io/en/0.5.3
~~ PlatformIO (recommended): platformio.org
~~ pybind11: github.com/pybind/pybind11
~~ libnpy: github.com/llohse/libnpy

In this video you will learn how to build a simple AI (inference only) project to send voice commands to a microcontroller over Wi-Fi.

We will use the wave2vec2 speech to text pre-trained model, export to ONNX, then perform inference in C++.

Next, we will create another C++ program to send the text to the microcontroller over TCP/WiFi.

Finally we will integrate the compiled modules into a Python script using pybind11.

Timestamps:
Setup: 00:01:17
Python Sandbox: 00:03:50
Inference Test (Py/CPP): 00:13:55
Building pybind modules: 00:22:34
Final Program: 00:36:15
Demo: 00:40:10

Thanks for watching!

~~My Udemy Courses on Motion Planning / Navigation / Flight Simulation / Aerospace Engineering
udemy.com/user/vinayak-deshpande-22
~~Buy Coffee: buymeacoffee.com/vdengineering
~~Instagram: instagram.com/vinayak_desh