Blog

I use this section to share technical notes related to my research and engineering work in multimodal AI. Current writing priorities include reasoning faithfulness in language models, noise-robust audio-visual speech recognition, and efficient speech tokenization.

Detailed posts will be published here as they are finalized.