We experimentally demonstrated different SKMs to perform machine learning analysis over complex, visible to mid-infrared (MIR) incident spectra, with the readout photocurrent directly providing the final inference. This process mathematically resembles the kernel machine algorithms typically used for digital machine learning, making each photodetector a spectral kernel machine. It can "sniff-and-seek," conceptually inspired by retriever dogs, learning from examples to recognize spectral features within a complex scene. We experimentally demonstrated different SKMs to perform diverse tasks, including visible-band image segmentation, nanoscale-thickness metrology of wafer oxide layers, discrimination of natural and artificial leaves, and identification of leaf hydration levels, achieving high accuracy under blind testing. For the MIR band, an electrically tunable bipolar photodiode made of bP-MoS2 heterostructure was operated at room temperature to identify chemicals and quantify mixture concentrations that were indistinguishable in the visible bands.