Explores the design of a general-purpose distributed execution system, covering challenges, specialized frameworks, decentralized control logic, and high-performance shuffle.
Explores model selection, evaluation, and generalization in machine learning, emphasizing unbiased performance estimation and the risks of over-learning.