Live siteAn explainable healthcare risk-modelling workflow built around model evaluation, experiment tracking, and a clinician-facing interface.
PythonLightGBMscikit-learnSHAP+6
AI/ML Developer
I'm Abdul Wahid, an Artificial Intelligence undergraduate building practical machine-learning systems, explainable workflows, and software around data.
Selected work
Not just a stack of technologies. Each case study explains the problem, the approach, the evidence, and what remains uncertain.
Explore all projectsLive siteAn explainable healthcare risk-modelling workflow built around model evaluation, experiment tracking, and a clinician-facing interface.
Live siteA supervised-learning study comparing classification approaches for identifying customers who may leave a service.
Live siteA course-resource web app for organizing lecture notes, PDFs, CDFs, and lab manuals with search and study helpers.
Live sitePrivate nearby file sharing designed to move files between devices without accounts, cloud uploads, or share links.
Working principles
My goal is to make work that can be read, questioned, and improved by someone other than its author.
Choose evaluation criteria that reflect the actual problem. A headline metric is a starting point, not a conclusion.
Show assumptions, weak spots, and the difference between a prototype and a validated system.
Connect notebooks and models to clear interfaces, reproducible workflows, and usable documentation.
Field notes
Short, research-informed notes on model evaluation, explainability, and responsible AI engineering.
A practical checklist for evaluating classifiers without hiding the trade-offs that matter.
Feature attribution is a lens on model behaviour, not a shortcut to causality or trust.
Project count reflects the currently documented portfolio, not an independent measure of impact. Details and limits are provided on each project page.
Open to conversations
For AI/ML roles, research opportunities, and thoughtful collaboration.