Vijayan Shinde

• Implemented Primal-Dual Algorithm to optimize GPU pricing under inventory constraints during off-peak hours. • Combined convex optimization with ML-based demand modeling across 2 inventory tiers. • Simulated 15,000+ realistic, stochastic demand scenarios using randomized logistic functions. • Achieved >85% prediction accuracy with Random Forest and LightGBM models for demand prediction. • Achieved near-oracle revenues with regret consistently <10% (vs. 15% benchmark).

Data Analytics and Software Development

Applied analytical methods in thermal simulations, FEA, metallurgy, and manufacturing optimization. Analyzed and designed quenching system for structural mild steel rebar using transient thermal and finite element analysis. Performed data-driven study correlating yield strength and elongation in structural steel under bending to optimize material performance.