Non-equilibrium activity is crucial for maintain and modulating tissue shape, development and morphogenesis, lysosome dynamics, cell membrane remodeling during cell division or membrane fusion and fission events. Importantly many of these processes play a significant role in human health helping regulate for instance immune system function and ensuring accurate developmental morphogenesis. However, there is a major gap in our understanding of how microscopic non-equilibrium biophysical driving forces give rise to a desired molecular response, function, or control. Indeed, while the theoretical and computational frameworks for the study of equilibrium biological processes are very well developed, there are very limited analogous tools for the study of complex far-from-equilibrium biological systems. Further, the large length and time scales of biological systems and processes make explicit computational simulations impractical. Addressing this problem requires the development of a range of multiscale non-equilibrium statistical mechanics techniques in combination with tools from machine learning and artificial intelligence so that the large length and time scales associated with the above-mentioned biological processes can be appropriately captured. The work outlined in this proposal builds towards these long-term goals by focusing on three paradigmatic example systems 1) Understanding and predicting non-equilibrium lysosomal dynamics and morphologies 2) Understanding and modelling cytoskeletal processes responsible for developmental patterning, cell-cell communication, and force generation 3) Developing frameworks for determining drivers of cell fate and differentiation from single cell RNA sequencing data. Each of these paradigmatic examples has implications for diseases. These paradigmatic examples build on the recent foundational non-equilibrium statistical mechanics frameworks developed by my group and expand them so that they can be utilized in biological contexts.