Work
Projects
Real builds — from ML models and data science to full-stack applications. Each one a step toward understanding a domain more completely.
Shipped
InboxSignal
Job Pipeline Tracker
A smart job application tracker that organises every stage of your pipeline — applied, screened, interviewed, offered. Clear visual status boards so nothing falls through the cracks.
CineMatch
Movie Recommendation Engine
Content-based movie recommender built using TF-IDF vectorization and cosine similarity. Enter a title and get curated suggestions ranked by feature overlap — genres, cast, and keywords.
VisionDigits
Digit Recognition Engine
A convolutional neural network trained on MNIST achieving 99%+ accuracy. Explores filter visualisation and activation maps to explain what the model is actually learning at each layer.
Classic Classifier Lab
Classification with SVM
A rigorous exploration of Support Vector Machines — kernel tricks, hyperparameter tuning with grid search, and boundary visualisation. A clean reference implementation for classical ML classification.
Falcon Landing Intelligence
SpaceX Landing Optimisation
Predictive modelling on SpaceX Falcon 9 landing outcome data. Applies logistic regression, decision trees, and random forests to understand which mission parameters most affect landing success.
GraphMind
Neural Graph Exploration
A deep-dive into Graph Neural Networks — node classification, link prediction, and graph-level tasks with GCN, GAT, and GraphSAGE architectures. Built to understand the theory through implementation.
Incoming Builds
Incoming Build
A full-stack product in active development. Details soon.
Incoming Build
Something in the data space. Architecture phase.