Grounding and Evaluation for Large Language Models: Practical Challenges and Lessons Learned (Survey)