Optimization of metabolic flux is a difficult and time-consuming process that often involves changing the expression levels of multiple genes simultaneously. While some pathways have a known rate limiting step, more complex metabolic networks can require a trial-and-error approach of tuning the expression of multiple genes to achieve a desired distribution of metabolic resources. Here we present an efficient method for generating expression diversity on a combinatorial scale using CRISPR interference. We use a modified native Escherichia coli Type I-E CRISPR-Cas system and an iterative cloning strategy for construction of guide RNA arrays. This approach allowed us to build a combinatorial gene expression library three orders of magnitude larger than previous studies. In less than 1 month, we generated similar to 12,000 combinatorial gene expression variants that target six different genes and screened these variants for increased malonyl-CoA flux and 3-hydroxypropionate (3HP) production. We were able to identify a set of variants that exhibited a significant increase in malonyl-CoA flux and up to a 98% increase in 3HP production. This approach provides a fast and easy-to-implement strategy for engineering metabolic pathway flux for development of industrially relevant strains, as well as investigation of fundamental biological questions.