- 👋 Hi, I’m @samsad35!
- 🎓 Postdoctoral Researcher in AI, with research spanning several key axes: Interpretability, Generative Models, Self-Supervised Learning (SSL), and Multimodal Data (with a special focus on audiovisual speech data).
- 🔬 My current work explores:
- Self-Supervised Learning: Advancing representation learning for speech and audio, including the development and evaluation of neural audio codecs.
- Interpretability: Analyzing latent spaces, preventing representation collapse, and building tools to evaluate complex representation metrics.
- Generative Models: Exploring modern generative paradigms for high-quality audio and speech modeling.
- Multimodal Architectures: Integrating multiple data streams, such as audiovisual and articulatory data, to build more robust predictive models.
- 🚀 Experienced in running large-scale distributed training on high-performance computing clusters (SLURM, multi-GPU environments).
- Languages: Python, HTML/CSS
- Machine Learning / AI: PyTorch, Self-Supervised Learning, Generative Models
- Audio & Signal Processing: Neural Codecs,
audiomentations, SSL Metric Evaluation - Infrastructure: SLURM (High-Performance Computing), Linux Environments
- Personal Website / Portfolio: samsad35.github.io