M.S. Data Science · Michigan Tech · AI/ML Specialisation
Building end-to-end intelligent systems that work in the real world - wherever the hardest problems are.
About
I'm a Master's student in Data Science (AI/ML specialisation) at Michigan Technological University. My foundation is a BSc in Information Technology with a specialisation in Software Engineering from Richfield Graduate Institute of Technology, Johannesburg, South Africa.
Before graduate school I worked across three countries — Zimbabwe, South Africa, and the US, building production systems as a data analyst and software engineer. I bring that production instinct to every ML project: the goal is always something that deploys, not just runs.
Open to collaborations across industries; if there's a hard problem and data, let's talk. Outside the lab: gym, dance, and travelling the world.
Education & Experience
Selected Work
From deployed production APIs to active research - spanning finance, healthcare, infrastructure, and construction AI. Click any project for the full case study.
End-to-end market stress prediction processing 870K+ financial headlines alongside FRED macro series. PyTorch Temporal Fusion Transformer, XGBoost, conformal prediction, HMM regime modeling. 9 live REST endpoints on HuggingFace Spaces via Docker. React dashboard on Vercel.
Unsupervised time-series deep learning for detecting abnormal behaviour in infrastructure load signals. Dense Autoencoder vs LSTM Autoencoder — learns normal patterns, surfaces deviations automatically. No labelled failure data required.
Multi-condition patient deterioration early warning system on MIMIC-IV clinical data. Designed to surface early risk signals across multiple clinical pathways simultaneously in ICU settings.
Agentic AI system combining computer vision, document intelligence, and autonomous reasoning pipelines for construction industry safety, compliance, and project intelligence.
Human Activity Recognition benchmark on UCI HAR smartphone sensor data across 6 activity classes. Full ablation study comparing CNN, LSTM, and hybrid CNN-LSTM architectures. Run on MTU's HPC cluster.
Data mining pipeline on CDC Chronic Disease Indicators. XGBoost, Random Forest, AdaBoost, KMeans, DBSCAN, PCA. Full cross-validation and cluster profiling for population-scale risk pattern analysis.
Statistical analysis of healthcare spending against population health outcomes across countries. Spearman correlation, OLS regression, and multi-panel visualisations examining spending-outcome relationships globally.
Full relational database design for an online store. ER modelling, normalised schema, complex multi-table SQL queries, stored procedures, and transaction management.
Multi-condition early warning system on MIMIC-IV. Real-time risk scoring across simultaneous clinical pathways for ICU settings.
Data mining on CDC CDI. XGBoost, Random Forest, KMeans, DBSCAN, PCA — population-scale chronic disease risk analysis.
Spearman correlation and OLS regression comparing healthcare spending to health outcomes globally.
Technical Stack
Languages
ML / AI Frameworks
Generative AI & APIs
Engineering
Data & Enterprise
Get in Touch
Available for full-time, internship, and collaborative roles in:
Also open to cross-disciplinary research collaborations and engineering partnerships across complex, data-driven domains.
slbimbi@mtu.edu