Random representation learning and feature selection for efficient classification without expensive training.
Clinical ML pipelines for respiratory analysis with uncertainty quantification and interpretable predictions.
GNN-based models for structured data, from CAD automation to autoregressive graph inference.
Hierarchical feature selection reducing 94% of features while preserving accuracy
IEEE MLSP 2024MCMC-based feature selection with LSH for time series classification
IEEE ICC 2023Random representation with Scatter Score and threshold exceedance rate
IEEE MLSP
I combine academic rigor with industry experience to build ML systems that work in the real world. Currently at UHN developing clinical tools, with a background spanning recommendation systems, IoT platforms, and distributed computing.
Read more about me