Selected Projects

CEL-Net: Collaborative Ensemble Learning

A peer-learning framework where neural networks teach each other via pairwise knowledge transfer during training to improve generalization under fixed compute budgets.

PyTorch, Knowledge Transfer, Ensembles

CPPN-Assisted Feature Distillation

Uses CPPNs as an evolutionary helper module to transform images into structured representations, enabling richer feature distillation for downstream CNNs.

Evolutionary Computation, PyTorch, Computer Vision

Generative AI Evaluation & Benchmarking

Tools and pipelines for benchmarking generative models with metrics and QA-style evaluations to improve reliability and deployment readiness.

Benchmarking, LLMs, Multimodal