AI-Guided Drug Delivery Parameter Prediction

AI-Guided Drug Delivery is a regression and ensemble modeling project focused on predicting optimal drug delivery parameters from biomedical simulation data. The work targets dosage, timing, release rates, and carrier geometry across multiple therapeutic classes, with an emphasis on safety, domain validity, and clinically relevant constraints. By embedding physics-informed rules directly into the model architecture, the system prioritizes plausible recommendations over raw score optimization.

A physics-informed ML pipeline that predicts safe, optimized drug delivery parameters across diverse biomedical scenarios.

Highlights

• —Developed regression and ensemble models to predict dosage, timing, release rates, and carrier geometry from simulation data across three therapeutic target classes
• —Integrated physics-informed constraints into the model architecture to enforce domain-valid outputs
• —Reduced out-of-bounds predictions by over 60% versus unconstrained baselines in internal validation
• —Designed evaluation protocols that balance accuracy metrics with clinical plausibility and safety-driven thresholds
• —Contributed to a research pipeline intended for publication and further collaboration with biomedical engineering teams