Research & teaching
Research with
engineering in mind.
I study how intelligent systems behave under imperfect data, adversarial conditions, and controlled multimodal changes — and help others learn the foundations behind them.
Ongoing research / Federated learning
FedMed-Defender.
Investigating Byzantine-robust federated aggregation under non-IID medical data and malicious client updates, using a PyTorch experimental framework.
My work includes implementing aggregation baselines, adversarial scenarios, and evaluation workflows while exploring robust statistics and update filtering.
Research collaborator
Advisor: Dr. Mohammadreza Shokournia
Ongoing research / Multimodal AI
Vision-language models
& interpretability.
Exploring how controlled image and question variations affect predictions and internal representations in vision-language models.
This involves exploratory multimodal experiments, hypothesis refinement, and evaluation design.
Research collaborator
Advisor: Prof. Mostafa Tavassolipour