The two tasks will use the same double-null, biased-down plasma shape. Though the high q|| task will need to push to high neutral beam injected power (Pinj) and plasma current (Ip), the early shots testing shape and x-point control are not sensitive to the injected beam power. The reinforcement learning (RL) controller tests are insensitive to shape, hence the early shape tests and the RL tests will run parallel in the first several plasma shots, testing overall shape, x-point location and control for the high q|| task while testing RL control with neutral beam injection as actuator for the RL control of ITB task. After completing R and Z scans of the x-point, the focus will shift to increasing the heat flux via increases in Ip and Pinj. There is an inherent tension between running at high Ip, moving the x-point to high Z, and moving it to high R. We will map out the optimal operating space in these parameters based on desired features of the plasma discharge: Ip>1.8 MA, attached divertor (which may require low Zeff and hence higher X-pt, and/or pumping and hence high X-point R), and steady Ip flattop with clear IR camera data for heat flux analysis.