Deepfake Voice Detection System

Overview

Telecom providers and financial institutions needed a solution to detect synthetic speech and protect against voice fraud in call centers. We built a real-time deepfake detection system for streaming audio using speaker recognition, diarization, ASR, Deep Fake detector model.

Achievements

The system was deployed in production, integrated into call center platforms, and used to flag synthetic audio segments and alert human operators in real time.

Responsibilities

• Collected and processed terabytes of real and synthetic audio.
• Created a modular training pipeline with automated KPI evaluation and CI/CD deployment to Triton.
• Built a streaming inference module with ensemble logic and GPU optimization.