Audio Recognition System

I engineered a local, privacy-focused audio recognition system (Shazam clone) from scratch.

Solution:

DSP Engine: Built a custom Node.js pipeline using FFT & spectrograms for noise-resilient fingerprinting.

Backend: Developed a high-performance Express.js API with optimized SQLite indexing.

Frontend: Created a real-time browser recorder for microphone and system audio.

Impact:

Achieved accurate song matching in <3 seconds.

Resolved critical recursion issues for 100% stability.

Delivered a fully portable, serverless database architecture.