With the shortage of rehabilitation clinicians in rural areas and the ongoing COVID-19 pandemic, remote rehabilitation (telerehab) fills an important gap in access to rehabilitation, especially for the treatment of adults and children experiencing upper arm disability due to stroke and cerebral palsy. We propose the use of a socially assistive humanoid robot coupled with a telepresence platform and computer vision based assessment pipeline as a tool for achieving more effective telerehab. The humanoid has arms, a torso, and a face to play games with and guide patients under the supervision of a remote clinician. The computer vision system is being developed to aid in automated objective assessments of patient function. In this talk, I will outline the design of such a system, present a uniquely large perceived usefulness evaluation of the platform with hundreds of practicing therapists in the United States, and present recent results from a pilot study with the system.