High performance and low cost sensors based on microelectromechanical systems (MEMS) have become commonplace in today's world. MEMS sensors, such as accelerometers, gy- roscopes, pressure sensors, and microphones, are routinely used in consumer electronics, automobiles, industrial and aerospace applications. Basically, all these devices mea- sure tiny displacements of micromachined mechanical structures in response to external stimuli. One of the widely used techniques to detect these displacements is piezoresistive sensing. Piezoresistive sensors are popular in MEMS due to their simplicity and robustness. Traditionally, silicon has been the material of choice for piezoresistors due to its high strain sensitivity or gauge factor. Whereas metal lm piezoresistors typically have low gauge factor that puts them out of favour when compared to silicon. But metal lm piezoresistors have several advantages compared to their semiconductor counterparts, including simple and low-cost fabrication, low resistivity and generally low noise. Low resistance sensors become desirable particularly when the devices are scaled down to nanoelectromechanical systems (NEMS), where signal-to-noise ratio (SNR) performance becomes crucial. Enhancing the gauge factor of metal lms while keeping their low resistance advantage can dramatically improve their SNR performance for NEMS. This thesis reports a simple method we have developed to enhance the gauge factor of metal lm piezoresistors. We demonstrate this method on specially designed micro- cantilever devices. Using controlled electromigration, we are able to engineer the microstructure of gold lm and transform it into a locally inhomogeneous conductor which resembles a percolation network. This results in more than 100 times higher gauge factor at low to moderate sensor resistance. The SNR possible with our piezoresistor at high frequencies exceeds that of most available systems by at least an order of magnitude. Our locally inhomogeneous metal lm piezoresistor is a promising candidate for high-performance NEMS-based sensors of the future.