Cancer is currently one of the leading causes of death worldwide with 1 in 6 deaths being cancer related. Conventional therapies such as surgery, radiation and chemotherapy have offered hope for patients and in many cases have resulted in complete tumor remission. Unfortunately, many tumors become resistant to conventional cancer therapies and require more effective targeting to achieve lasting remissions or cures. Immunotherapies, such as monoclonal antibodies and chimeric antigen receptor (CAR) T cells, exert substantial anti-tumor activity and have advanced the field greatly. However, due to their high specificity for a single target antigen, these therapies generally lack tumor selectivity resulting in on-target, off-tumor toxicity against healthy tissues. The central focus of my dissertation is to develop immunotherapies that are highly selective for tumor cells and spare healthy tissues. To prevent on-target, off-tumor toxicities, three separate approaches were taken with different therapies. The first approach focused on the generation of low-affinity CAR constructs using systematic mutagenesis of a parental high-affinity single-chain variable fragment (scFv) to more precisely target CD229 on multiple myeloma cells, while sparing healthy CD229-positive lymphocytes. The second approach focused on restricting the trafficking of LINGO1 CAR T cells through integrin knockout. LINGO1 CAR T cells with restricted trafficking would be able to target LINGO1-positive Ewing sarcoma cells but not LINGO1-positive healthy cells in the brain. The third approach taken to engineer tumor selectivity focused on the development of sweeping antibodies against interleukin 16 (IL-16), an antigen that is soluble, pro-tumorigenic and has minimal physiological function. The significant findings from these three studies are: (1) low-affinity CAR constructs can be generated from a high-affinity CAR construct and spare healthy T cells while maintaining anti-tumor activity. (2) Integrin knockout in activated, primary T cells prevents binding to VCAM1 and trafficking of CAR T cells to the brain. (3) Sweeping antibodies against IL-16 could be generated and clear soluble IL-16 in vitro. Together, these approaches show that selectivity can be engineered into both monoclonal antibodies and CAR T cell therapies to reduce on-target off-tumor toxicity and increase anti-tumor efficacy.