Particle tracking of microscopic species is widely employed to infer biophysical transport properties and to validate prospective species transport models. “Particles” include viruses, bacteria, bacteriophages, passive or active microbead probes, and drug carrier nanoparticles. The major bottlenecks in application of predictive mathematical models to micron-scale living systems are low data throughput (manual conversion of videos to an ensemble of position time series) and the inability of 2D particle tracking to capture rapidly diffusing, low SNR species for sufficiently long duration. I will present a machine learning algorithm that automates the conversion of 2D and 3D raw videos and makes feasible, for the first time, 3D particle tracking of sub-micron species at low SNR, for which 2D particle tracking has been futile. The artificial neural network enables, for the first time, sufficient data for Bayesian model selection and parameter estimation of these important transport processes. I will discuss two ongoing projects that use automated particle tracking: (1) Flagellar-mediated active motion of Salmonella in visco-elastic biogels; and (2) Particle tracking in living fungal cells, using genetically engineered multimeric probes.