Physics-informed learning has been extensively used across various domains, but its computational efficiency and practicability in engineering practice often incur scepticism. Particularly, its application in geotechnics remains in its infancy. To answer these questions, we enhance current physics-informed learning by integrating adaptive sampling, domain decomposition, Ritz method, transfer learning and customised optimizer and loss functions. Its feasibility is demonstrated by applying it to non-linear time-dependent 1D and 2D consolidation equations, elastic and elasto-plastic cavity expansion and footing problems. The results indicate that enhanced physics-informed learning enables solving time-dependent, elastic and elasto-plastic problems, whilst using transfer learning can achieve a significant acceleration in computation. This framework allows for flexible data assimilation and its combination with transfer learning enables the utilization of multi-source data, including sparse in-situ monitoring and historical data, to derive accurate solutions for problems at hand. This capability offers a promising and cost-effective strategy for sensor installation, real-time prediction, model parameters identification and digital twinning in engineering practice.