Research

For a complete and frequently updated bibliography, please see my Google Scholar. I keep this page as a compact research summary rather than a duplicate publication database.

Last manually curated: August 2026.

My research studies the two-way interaction between optimization, operations research, and artificial intelligence. My established work develops optimization principles and algorithms for building modern AI systems. I am also developing a complementary research direction that uses AI to improve optimization algorithms and operational decision-making.

Optimization and Operations Research for AI

I develop optimization theory and algorithms for training and aligning modern machine-learning systems, especially foundation models. My goal is to connect mathematical principles with the computational, memory, communication, and data constraints encountered in practice.

Current themes include:

• Optimization algorithms and training dynamics for foundation-model pre-training
• Reinforcement learning and optimization for post-training/alignment
• Memory- and parameter-efficient pretraining and fine-tuning
• Bilevel, minimax, and multi-objective learning
• Distributed, decentralized, and federated optimization
• Riemannian and zeroth-order optimization

AI for Optimization and Operations Research

I am developing methods that use AI—particularly foundation models and learning-augmented algorithms—to improve how optimization and operations research problems are formulated, solved, and deployed.

I am especially interested in questions such as:

• How can learned models identify useful structure in optimization problems?
• How can AI guide algorithm selection, configuration, and search?
• How can foundation models assist with optimization modeling and operational decision-making?
• How can learned components be integrated with optimization algorithms while preserving reliability and performance guarantees?

Recent Highlights

• Towards joint scaling laws with optimal batch size schedules. Preprint, 2026 [PDF]
• A minimalist optimizer design for LLM pretraining. ICML 2026 [PDF]
• A Tale of Two Problems: Multi-Task Bilevel Learning Meets Equality Constrained Multi-Objective Optimization. ICML 2026
• Muon Outperforms Adam in Tail-End Associative Memory Learning. ICLR 2026 [PDF]
• Joint Demonstration and Preference Learning Improves Policy Alignment with Human Feedback. ICLR 2025 (Spotlight) [PDF]

Selected Publications

• Riemannian Bilevel Optimization. Journal of Machine Learning Research [PDF]
• A Riemannian ADMM. Mathematics of Operations Research [PDF]
• Zeroth-order Riemannian Averaging Stochastic Approximation Algorithms. SIAM Journal on Optimization [PDF]
• Problem-Parameter-Free Decentralized Nonconvex Stochastic Optimization. Pacific Journal of Optimization [PDF]
• Stochastic Zeroth-order Riemannian Derivative Estimation and Optimization. Mathematics of Operations Research [PDF]
• Getting More Juice Out of the SFT Data: Reward Learning from Human Demonstration Improves SFT for LLM Alignment. NeurIPS 2024 [PDF]
• Revisiting zeroth-order optimization for memory-efficient llm fine-tuning: A benchmark. ICML 2024 [PDF]