This talk presents a control-oriented perspective on Scientific Machine Learning (SciML) for modeling, optimization, and control of dynamical systems. SciML provides a unifying computational paradigm that integrates physics-based models, optimization algorithms, and control policies within a differentiable programming framework. This synthesis enables computation of structured gradients for constrained system identification, constrained optimization, and learning-based control problems while preserving interpretability, stability, and physical consistency. Two recent advances will be highlighted. First, differentiable predictive control [1], a SciML approach that merges model predictive control with gradient-based learning to enable scalable, self-supervised training of explicit control policies suitable for real-time deployment on embedded hardware. Second, an operator-splitting formulation for neural differential-algebraic equations [2] that integrates mechanistic dynamics with neural components to achieve robust extrapolation in systems with implicit constraints and conservation laws. Together, these advances demonstrate how SciML can unlock new capabilities for the modeling, optimization, and control of complex dynamical systems, with applications in power grid and building energy management. References [1] J. Drgoňa, A. Tuor and D. Vrabie, "Learning Constrained Parametric Differentiable Predictive Control Policies With Guarantees," in IEEE Transactions on Systems, Man, and Cybernetics: Systems, vol. 54, no. 6, pp. 3596-3607, June 2024 [2] James Koch and Madelyn Shapiro and Himanshu Sharma and Draguna Vrabie and Ján Drgoňa, Learning Neural Differential Algebraic Equations via Operator Splitting, Conference on Decision and Control (CDC), 2025.