Predictive Machine Learning for Industrial Coatings

Through an IISc–Accenture collaborative initiative, I helped develop a predictive machine learning framework for next-generation industrial paint formulations. This involved building high-fidelity molecular models of ternary systems — Poly(methyl methacrylate) (PMMA), Polystyrene (PS), and TiO₂ nanoparticles — to study interfacial physics and dispersion stability, and architecting a data-driven model mapping molecular-level descriptors (polymer architecture, grafting density, nanoparticle loading) to macroscopic rheological and mechanical properties such as viscosity, pigment dispersion quality, and film formation.