We present a camera-in-the-loop reinforcementlearning (RL) approach to MIMO beam steering with safety aware rewards. The pipeline logs reward curves and producesθ–f heatmaps for learned beams using lightweight scripts wiredto a Makefile. bgilbert1984