Add documentation and example code to: - create a new model/agent - manage a new JSON dataset - run an experimental pipeline through slurm - check output and metrics - run inference on a trained model