AI Data Labeler & Model Auditor — Deepfake Detection Project
OtherImageClassification
Trained GAN-based models to generate high-fidelity synthetic images and detect deepfake content through AI-driven classification. Developed feature classification protocols for differentiating AI-generated and organic media, contributing directly to LLM output auditing. Evaluated anomaly detection procedures aiding in refining model labeling for AI safety and adversarial robustness. • Prepared and labeled datasets of synthetic and real images for classification tasks.
• Tuned feature extraction and anomaly detection methods for improved accuracy.
• Implemented model-driven quality checks for dataset integrity.
• Employed StyleGAN, FOMM, PyTorch, TensorFlow, OpenCV, and Colab in image data labeling.
Trained GAN-based models to generate high-fidelity synthetic images and detect deepfake content through AI-driven classification. Developed feature classification protocols for differentiating AI-generated and organic media, contributing directly to LLM output auditing. Evaluated anomaly detection procedures aiding in refining model labeling for AI safety and adversarial robustness. • Prepared and labeled datasets of synthetic and real images for classification tasks. • Tuned feature extraction and anomaly detection methods for improved accuracy. • Implemented model-driven quality checks for dataset integrity. • Employed StyleGAN, FOMM, PyTorch, TensorFlow, OpenCV, and Colab in image data labeling.
2025 - 2025