This coursera class on Bayesian Methods for Machine Learning by Daniil Polykovskiy and Alexander Novikov is an excellent introduction to key Bayesian topics that are not always well understood by people such as

  • Variational Inference
  • MCMC
  • Latent Variable Models
  • Topic modeling
  • Reparameterization trick
  • Variational autoencoders
  • Variational dropout
  • Gaussian processes
  • etc.