With the availability of cost-effective digitizers and powerful pipeline processors, modern spectroscopy can be performed completely in the digital domain with minimal analog front-end signal processing.  These systems offer the flexibility and extensibility of digital signal processing algorithms to simultaneously extract multiple waveform parameters such as pulse height and start time from faint signals buried in background noise.  However, pulse fitting was traditionally a computationally intensive task performed offline on computing clusters in the final analysis. I present new algorithms we have developed to perform full nonlinear covariant least-squares fits for optimal resolution of 30 keV proton pulses in the Nab experiment.  These algorithms can be implemented in digitizer firmware to increase the resolution and efficiency of our hardware trigger as well as on graphics processing units (GPUs) for high precision online analysis of triggered waveforms.