Data Scientist | AI Researcher | Mathematician
Interested in
|
I am a researcher passionate about building efficient, interpretable, and robust AI systems.
My work focuses on neural network pruning, data efficiency, and
explainability in deep learning. I enjoy connecting theory with application,
particularly in domains like vision, language, and
graph learning.
I enjoy blending theory with real-world applications, whether it's making neural networks smarter and faster, adapting models across domains, or explaining why a model made its decision. I'm currently in the U.S.
and always looking to connect with collaborators or teams passionate about advancing AI.
Machine Learning, Deep Learning, Data Science, Neural Network Optimization, Data Efficiency, Robust Nueral Network, Continual Learning, Network Interpretability, Natural Languge Processing, Graph Neural Network, Graph Theory and Applications, Combinatorics, Probability, Mathematics
My recent projects in deep learning, continual learning, and neural network interpretability.
We investigate using synthetic data (GANs, Diffusion) to replace natural samples in continual task learning. Our EpochLoss strategy removes uninformative examples, enhancing adversarial robustness and model generalization.
A theoretical framework to explain DNN pruning through information-theoretic divergence in latent spaces. Introduces novel projection metrics AP2 and AP3 and validates them on ResNet, VGG16, and ViT.
Ph.D. Student in Computer Science. Research focuses on deep learning model interpretability and optimization, aiming to make neural networks more compact, robust, and explainable.
GPA: 3.953 / 4.00
Researcher (Herz Fellow), Alexander von Humboldt Foundation Grant. Worked on graph algorithms and optimization methods in theoretical computer science.
Research assistance in Mathematics. Admitted as a Distinguished Talent. Conducted research in extremal graph theory with a focus on graph coloring problems.
GPA: 3.96 / 4.00
M.Sc. in Mathematics. Ranked 2nd in M.Sc. Specialized in graph coloring, combinatorics, and their applications in theoretical computer sciences.
GPA: 3.91 / 4.00
B.Sc. in Mathematics. Ranked 4th. Built strong foundations in algebra, Probability, discrete mathematics, and graph theory.
GPA: 3.63 / 4.00
Developed robust and efficient CL frameworks using synthetic data and sample removal techniques like EpochLoss.
Applied prototypes to both regression and classification, enabling token/patch-wise interpretability in text and vision models.
Focused on receptive fields, prototype visualization, and decision tree–based interpretation for explaining model behavior.
Proficient in PyTorch and scientific Python libraries for building, training, and debugging scalable neural network pipelines.
Feel free to reach out to me for collaborations, research discussions, or any questions. I am always happy to connect!