ML benchmarking framework for climbing route prediction. Designed a reproducible evaluation protocol with 5-fold retrain-per-fold cross-validation, tested 8+ model architectures (MLP, LSTM, CNN, Transformer, Ridge, Perceptron), and established a non-tree leaderboard. Built dataset augmentation pipeline with hold-swap tokenisation and identified/patched data leakage bugs.
A novel framework in Python with PyTorch for evaluating neural networks in harsh environments. Proposed two novel activation functions that proved to be more robust to SEE related errors.
Developed novel time-series features for classification of photoplethysmography signal segments. Tested a range of ML architectures and hyperparameters using AUC+ROC metrics. Manuscript prepared but not submitted for publication.