NLP From Scratch

I developed an NLP neural network in Python and PyTorch and integrated it into a full-stack TypeScript app today.

🏗️ I built a proof of concept for demonstrating how a natural language processing (NLP) model can output origin classification for a given first or last name based on a large dataset of names, a recurrent neural network (RNN), training the RNN on the dataset, evaluating the given name with the trained RNN, and then visualize a confusion matrix as a heat map for correct guesses of the name’s origin. Lastly, I exposed the RNN model to an Express.js API and consumed it in an Expo app.

🧑‍🍳 Ingredients:
• PyTorch
• NumPy
• Matplotlib
• Python
• Makefile
• Shell
• Express.js
• Node.js
• Expo
• Expo Router
• React
• React Native
• TypeScript
• Tailwind CSS
• NativeWind

🧑‍💻 Code available here:
https://github.com/pjnalls/expo-for-architects