Research Engineer

Sravanth Kodavanti

Research Engineer at Nava

I make AI models run fast and efficiently on GPUs. At Nava I work on inference optimization for large models; before that I spent two years at Samsung Research compressing LLMs, diffusion and image-restoration models for on-device deployment through neural architecture search, quantization, pruning and distillation. My work spans both the vision and language domains, with 10+ papers at venues like CVPR, ACL and ICASSP and two patents filed.

• CV