Cells need to be able to precisely control their proteome composition in order to maintain homeostasis. This is of particular relevance in fast dividing cells, such as embryonic stem cells, since their global protein levels need to be doubled at each division to ensure reliable growth and pluripotency maintenance. However, how protein synthesis and degradation interplay to regulate global protein expression levels throughout the cell cycle remains unclear. In addition, while cell-to-cell variability at the level of transcription and protein expression is relatively well studied, it remains unknown to what extent protein degradation rates vary between single cells. Previous studies focusing on protein turnover have mainly relied on bulk biochemical approaches, thus only providing population averages and a low temporal resolution. In addition, many approaches to study the changes in protein turnover throughout the cell cycle have been dependent on cell cycle synchonizing drugs, which might interfere with protein synthesis and also reduce cell viability. Therefore, we further developed and applied two fluorescence time-lapse imaging strategies to study the cell cycle dependence and cell-to-cell variability of protein synthesis and degradation in unperturbed dividing mouse embryonic stem (ES) cells: 1) a Mammalian Cell-optimized Tandem Fluorescent Timer (MCFT), consisting of the fusion of the fast maturing green fluorescent protein superfolder GFP (sfGFP) and the slower maturing red fluorescent protein mOrange2, which in combination with computational modeling allows for disentangling changes in protein synthesis from changes in protein degradation in single living cells, and 2) fluorescent pulse labeling of protein tags (SNAP/Halo-tags) to monitor protein half-lives, as well as transient changes in protein degradation rates in single cells without the use of protein synthesis inhibitors. We discovered a substantial cell cycle-dependence in both protein synthesis and degradation. In particular, a large fraction (40%) of proteins was stabilized around cytokinesis, an effect that has not been described previously. Interestingly, protein degradation rates were highly variable between single cells and were positively correlated with synthesis rates in individual cells to buffer protein expression level variability. Furthermore, degradation rates of different proteins co-varied between single cells, both in ES cells and fibroblasts, suggesting variability in proteasome activity as a source of global extrinsic noise in gene expression. We identified the proteasome component ADRM1 as a potential limiting factor that contributes to generating such noise, as its expression levels were negatively correlated with protein half-lives in single cells. Our work demonstrates the importance of single cell live imaging approaches to study dynamic processes and cell-to-cell variability and sheds more light on the complex interplay of protein synthesis and degradation in determining protein expression levels in single mammalian cells.