Portfolio
Selected Projects
A selection of projects across machine learning, biomedical data, neuroscience, and applied research tooling. Each one reflects the kind of work I enjoy most: technically rigorous, practically useful, and grounded in real questions.
Uncovering ADHD and Gender Disparities Using Brain Imaging and Machine Learning
Built a multi-outcome modeling workflow for the WiDS Datathon 2025 to predict ADHD diagnoses and explore sex-based differences using brain imaging and behavioral data. The project focuses on equitable model performance and the underdiagnosis of ADHD in females.
View project
SimSiam: Self-Supervised Feature Learning with PyTorch
Implemented and documented SimSiam for learning meaningful visual representations without labels. The project covers the training pipeline, evaluation flow, and visual analysis of learned features on CIFAR-10.
View project
Exploring Variational Autoencoders
Developed a hands-on VAE implementation on the SVHN dataset, covering model construction, training, and result visualization. This project explores how generative models can learn compact latent representations of image data.
View project
Connectivity Analysis with MVAR and ICA
Applied multivariate autoregressive modeling and independent component analysis to EEG time-series data in order to estimate transfer functions and directed coherence measures for connectivity analysis.
View project
Sentiment Analysis with Machine Learning and LIME
Built a sentiment classification pipeline using Naive Bayes, SVMs, and neural networks, then used LIME to make model predictions easier to interpret. The project combines practical text classification with explainable AI methods.
View project
Data Harmonization with R
Focused on aligning and integrating datasets to improve consistency and usability for downstream analysis. This project reflects an important part of practical machine learning work: making data trustworthy before modeling begins.
View project