Additive manufacturing (AM) has brought about a new age of design and manufacturing where traditional constraints are eased or in some areas completely lifted. The integration of topology optimization (TO) and AM has the potential to revolutionize modern design and manufacturing. However, few instances of manufactured optimized designs are documented, and even fewer examples of experimentally tested designs are available. The lack of validation combined with the influence of AM process on material properties leaves a gap in our understanding of process-microstructure-property relationships that is essential for developing holistic design optimization frameworks. In the first chapter of this thesis, a functional design was topologically optimized and fabricated using both directed energy deposition (DED) and selective laser melting (SLM), also known as powder bed fusion (PBF), methods. It is shown that TO results are sensitive to the AM method, post-processing conditions, and differences in mechanical properties. Thus, a TO for AM framework can be best optimized with the incorporation of microstructure features to account for localized microstructural variations in fabricated components. One of the main challenges facing SLM process is finding suitable process parameters to achieve maximum density (pore-free) parts. In chapter two, two newly discovered dimensionless numbers are presented that correlate process parameters to a part’s density allowing for an initial estimation of suitable process parameters without the need for extensive modeling or experimentation. The dimensionless numbers allow for identification of process parameters that will result in a maximum density regime in the as-built part. Finally, a universal scaling law is introduced that can aid in quantitative prediction of process parameters that result in the highest as-built density. Multi-material additive manufacturing using PBF has the potential to revolutionize the manufacturing landscape by producing parts with improved thermophysical properties and enhanced functionality. However, the complex nature of the process requires a case-by-case approach to multi-material characterization. The aim of chapter three is to provide firsthand knowledge of 316L stainless steel (316L) and Hastelloy X (HX) multi-material processing via PBF. Results showed that using the proper process parameters for each individual material led to formation of a compositional gradient at the interface that stretched for 240 µm (10-12 layers) with no evidence of cracking or porosity. Finally, it was concluded that the “naturally” formed interface created a compositional gradient which was defect free due to similar values of coefficient of thermal expansion, energy density, dimensionless number Π2, and different Marangoni numbers of the materials. The aim of chapter four is to propose a methodology for rapidly predicting suitable process parameters for additive manufacturing of multi-material parts with a compositional gradient by using machine learning and 316L-Cu bi-metal system. Specifically, an algorithm based on a multivariate Gaussian process is developed to predict part density and surface roughness for a given set of laser power, velocity, and hatch spacing values. After the model is validated using leave-one-out cross validation method, process parameter maps are generated for 316L-Cu parts manufactured using selective laser melting with premixed powder at mass fractions of 0.25, 0.50, and 0.75. It is shown that process parameters are a nonlinear function of gradient composition and neither process parameters of 316L or Cu are suitable for the graded region of a 316L-Cu multi-material part. Generated process maps provide a firsthand knowledge of process-property relationships for regions of compositional grading in 316L-Cu parts.