Souvik Kundu

Mixture of Agents

Reasoning Models

Inference Efficiency and Optimizations for LLMs

Long context Understanding and Reasoning of LLMs

Mixture of Experts

Pruning of DNN Model for Computer vision and NLP

Federated Learning for Resource-constrained Edge

Model and Data Privacy in Deep Learning

Brain-Inspired Spiking Neural Network Training and Compression Algorithms

Efficient Training of Robust DNNs against Adversarial Attacks

Machine Learning Architectures