Resonant optical sensors have enabled the label-free measurement of nanoparticles suspended in liquids, down to the resolution of individual viruses and large molecules, but are only able to quantify optical properties (refractive index, scattering, fluorescence). Additionally, these sensors are fundamentally limited by the random diffusion of particles to the sensing region, and thus only quantify a tiny fraction of the analyte. We have developed a microfluidic optomechanical resonator capable of sensing flowing nanoparticles using the action of phonons that are coupled to light. The phonon mode of the system casts a nearly perfect net for measuring density, viscoelasticity, and compressibility of the particles that flow through, without being limited by random diffusion. Information on the particle mechanical properties is encoded in the light scattered from the thermal fluctuations of the phonon mode, and measurements at a timescale of below 20 milliseconds have been demonstrated previously. In this work, we develop a new experimental method for improving the signal-to-noise ratio (SNR) and sensing speed achievable with this technique, by implementing electro-opto-mechanical transduction. We demonstrate real-time particle transit measurements as fast as 400 microseconds, a factor of 50x improvement in speed, without any post-processing. We discuss how this novel technique can be used for ultra-high throughput analysis of mechanical properties of biological particles in liquids, enabling a new form of flow cytometry.