The interplay of smart light bulbs (equipped with wireless controllable LEDs) and mobile sensors (embedded in wearable devices, such as smart watches and spectacles) enables a wide range of interactive lighting applications. One notable example is a smart lighting control system that provides automated illuminance management by wearable sensors close to end-users. In this paper, an energy-efficient smart lighting control system is developed using mobile light sensors for measuring local illuminance and assisting smart light bulbs to coordinate the brightness adjustments, while meeting users' heterogeneous lighting preferences. A pivotal challenge in these systems is attributed to the presence of oblivious mobile sensors hampered by the uncertainties in their relative locations to light bulbs, unknown indoor environment and time-varying background light sources. To cope with these hindrances, we devise an effective model-agnostic control algorithm inducing continuous adaptive coordination of oblivious mobile sensors without complete knowledge of dynamic operational environment and the associated parameters. The proposed algorithm is corroborated extensively under diverse settings and scenarios in a proof-of-concept smart lighting testbed featuring programmable light bulbs and smartphones, deployed as light sensing units. Lastly, we discuss some practical limitations of the proposed control approach, along with possible solutions, and conclude by outlining promising directions for future work.