Precision oncology seeks to exploit tumor-specific drug sensitivities. Traditionally, this is accomplished through the identification and targeting of highly recurrent mutations. This paradigm falls short in ovarian cancer where the oncogenic alterations are more diverse, necessitating an alternate approach for the identification of tumor-specific vulnerabilities. To address this, we have used a functional modeling approach, integrating drug screening with a Kinome Atlas-based assessment of signaling, to nominate a therapeutic regimen for ovarian tumors. This approach identifies a small-molecule RAS mimetic, rigosertib, as a tumor-selective agent and leads us to identify the combination of rigosertib with phosphoinositide 3-kinase (PI3K) or mammalian target of rapamycin (mTOR) inhibition as effective combinations that prevent rigosertib-induced survival signaling while inducing regressions in ovarian cancer xenografts. These data support further exploration of these combinations for the treatment of ovarian cancer.