Serverless computing, also known as Function-as-a-Service (FaaS), has gained significant interest in both academia and industry. Today many cloud service providers offer serverless platforms, including Amazon Web Services (AWS Lambda), Microsoft Azure (Azure Functions), and Google Cloud (Google Cloud Functions). Serverless computing empowers developers to build and manage applications without the need for provisioning or managing servers or backend infrastructure. However, the distinctive programming model and runtime environment inherent to serverless computing present unique challenges when developing new microservices applications or migrating existing ones to serverless platforms. This dissertation introduces several systems, namely Beldi, Mu2sls, MuCache, and CausalMesh, designed to address those issues.