AI for Quantum Computing

- Lead a research project: design and implement a large language model serving system - Optimize batching algorithm based on workload prediction and integer linear programming

- Design and evaluate ML models to predict GPU idle pattern- Design and implement a pipeline for automatic GPU idle prediction based on GPT-4

- Participate in a research project: benchmark and evaluate machine learning interpretability tools, where I test and benchmark 7 feature attribution tools- Weekly AI safety reading group

- Lead the StarLeap Project between Microsoft Research Asia and Microsoft Research Redmond: Intelligent Power Aware VM Allocation-VM allocation algorithm design, implementation, and evaluation with large scale production trace data

- Help developing an infrastructure operation system based on micro-services- CI/CD and documentation