Selected research work and projects demonstrating both theoretical understanding and practical implementation.
Reproducing and extending Wenkel et al.'s GCON architecture for graph combinatorial optimization problems including Maximum Cut, Maximum Clique, and Minimum Dominating Set. Explores the role of hybrid filter banks and localized attention mechanisms in unsupervised learning settings.
Simulation-based exploration of bias–variance tradeoffs and noise effects in linear models. Focus on statistical assumptions and interpretability.
Designing data ingestion, transformation, and analysis pipelines under real-world scalability and reliability constraints.