The recent surge in inflation has drawn increased attention to how central banks monitor economic conditions and calibrate policy responses. This study implements a Hemisphere Neural Network whose architecture allows the estimation of the unobserved states within the New Keynesian Phillips Curve and Taylor rule frameworks. Applying this model to Philippine macroeconomic data, we find that the HNN-derived latent states capture key economic concepts with outputs that align with well-known historical events. Long-run inflation expectations appear anchored at 3.5–4.5 percent, while commodity prices explain most short-term inflation spikes. The estimated output gap and inflation expectations are also broadly consistent with benchmark data from the Bangko Sentral ng Pilipinas (BSP) and uncover key features in inflation dynamics such as the flattening of the Phillips curve and the presence of diagnostic expectations. Extending the approach to a Taylor Rule setting also reveals insights into the BSP’s policy behavior and its alignment with latent economic conditions. Overall, BSP’s policy actions appear to have largely adhered to the standard principles of optimal monetary policy. Finally, the research contributes both an empirical tool and a modeling framework for monitoring policy-relevant inflationary pressures and supports the formulation of effective, data-informed monetary policy.