ML Models Delivered
A deeper look at delivered models and end-to-end pipelines (problem → data → approach → output).
Islamic Finance Screening
Problem: classify company revenues/activities as halal/haram compliant.
Data: SEC/EDGAR filings + financial datasets from LSEG (Refinitiv).
Output: structured JSON (revenue breakdown, business activities, compliance labels) via a real-time REST API.
CVD Genetic Deviations
Problem: identify genetic patterns and deviations associated with cardiovascular disease.
Approach: data preprocessing + feature engineering + ML modeling and interpretation for clinically meaningful signals.
Output: analysis pipeline and model results used for biomarker discovery.
Kidney Transplant Outcomes
Problem: model transplantation efficiency/outcomes using clinical and research data.
Approach: supervised ML with careful validation and feature analysis.
Output: predictive model and insights to support decision-making.
Diabetes & Peripheral Artery Disease
Problem: analyze genetic and clinical signals in diabetes with peripheral artery disease.
Approach: ML-driven analysis with focus on interpretability and biomedical relevance.
Output: analysis reports and candidate biomarkers.
Single-Cell RNA + Somatic Mutations (Grant Work)
Problem: research somatic mutations using single-cell RNA datasets; extract patterns and support diagnostic/therapy biomarkers.
Approach: end-to-end pipeline from genetic preprocessing and QC to ML analysis and reporting.
Output: reproducible analysis workflow and results for research stakeholders.
Want deeper technical details (metrics, datasets, stacks)? Tell me which models to expand first.