The lost foam casting (LFC) process is emerging as a reliable casting method. The metal-fill profile in LFC plays an important role among several factors that affect casting quality. The metal-fill profile is in turn affected by several factors. Several casting defects may result due to an improper metal-fill process. Hence, it becomes essential to characterize and control, if possible, the metal-fill process in LFC. This research presents instrumentation and a technique to monitor and characterize the metal-fill process. The characterization included the determination of the position of the metal front and the profile in which the metal fills up the foam pattern. The instrumentation included capacitive sensors. Each sensor is comprised of two electrodes whose capacitive coupling changes as the metal fills the foam pattern. Foundry tests were conducted to obtain the sensors' responses to the metal fill. Two such sensors were used in the foundry tests. Data representing the responses of these sensors during the metal-fill process were collected using a data acquisition system. A number of finite element electrostatic simulations were carried out to study the metal-fill process under conditions similar to those experienced in foundry tests. An artificial neural network was trained using the simulation data as inputs and the corresponding metal-fill profiles as outputs. The neural network was then used to infer the profile of the metal-fill during foundry tests. The results were verified by comparing the metal-fill profile inferred from the neural network to the actual metal-fill profile captured by an infrared camera used during the foundry tests. The match up between the inferred profiles and the infrared camera measurements was satisfactory, indicating that the developed technique provides a reliable and cost effective method to monitor the metal-fill profile in LFC.