Yifei Ming

Contact: alvinming5 [at] gmail [dot] com

Hi! I am a senior research scientist at Google. I work on research that advances Gemini’s agentic capabilities on open-ended, long-horizon, and economically valuable tasks, along with the ecosystem that supports them, such as agent harness, environment scaling, and self-evolution.

In the past, I worked on:

  • Post-training: training agents for tool use, long-context reasoning, deep research, and multi-turn interaction

  • Foundational research: understanding the capabilities and limits of LLM-based systems

  • Evaluation & inference scaling: building sandboxes, challenging benchmarks, and improving agentic workloads at test time

I am open to collaborations. We are also hiring self-motivated student researchers / research interns. If you are interested in the directions below, feel free to reach out.

Research directions

(1) Self-evolving agent. Agents can self-evolve by learning from past experience, but today’s trial-and-error process is largely heuristic-driven, time-consuming, prone to overfitting, and poorly supported by existing systems. What infrastructure and algorithms can support a continuously evolving agentic system?

(2) Long-horizon agent. Economically valuable tasks often require humans to work for a long time. When agents operate at that scale, they tend to fail in ways humans wouldn’t (e.g., poorly followed instructions, reward hacking, and context pollution), leading to saturating returns. How do we design scalable environments and agent harnesses that hold up over long horizons?

News

05/2026 Starting a new position as a Senior Research Scientist at Google
05/2024 Starting a new position as a Research Scientist at Salesforce
05/2024 Defended my Ph.D. thesis on Reliable Foundation Models in the Open World :mortar_board:
06/2023 Research intern at Microsoft Research working on spatial and mathematical reasoning for vision-language models.
05/2022 Research intern at Adobe working on multi-modal document understanding and robustness.
More News

Recent publications

  1. EMNLP
    Harnessing LLM Agents with Skill Programs
    Hongjun Liu, Yifei Ming, Shafiq Joty, and Chen Zhao
    In Conference on Empirical Methods in Natural Language Processing (EMNLP) 2026
    Paper
  2. ICML
    MAS-Orchestra: Understanding and Improving Multi-Agent Reasoning Through Holistic Orchestration and Controlled Benchmarks
    Zixuan Ke, Yifei Ming, Austin Xu, Ryan Chin, Xuan-Phi Nguyen, Prathyusha Jwalapuram, Jiayu Wang, Semih Yavuz, Caiming Xiong, and Shafiq Joty
    In International Conference on Machine Learning (ICML) 2026
    Code Paper
  3. ACL
    Hard2Verify: A Step-Level Verification Benchmark for Open-Ended Frontier Math
    Shrey Pandit, Austin Xu, Xuan-Phi Nguyen, Yifei Ming, Caiming Xiong, and Shafiq Joty
    In Annual Conference of the Association for Computational Linguistics (ACL) 2026
    Paper
  4. ACL
    Discovering the Gems in Early Layers: Accelerating Long-Context LLMs with 1000x Input Token Reduction
    Zhenmei Shi, Yifei Ming, Xuan-Phi Nguyen, Yingyu Liang, and Shafiq Joty
    In Annual Conference of the Association for Computational Linguistics (ACL) 2026
    Paper
  5. ICLR
    LiveResearchBench: A Live Benchmark for User-Centric Deep Research in the Wild
    Jiayu Wang*, Yifei Ming*, Riya Dulepet, Qinglin Chen, Austin Xu, Zixuan Ke, Frederic Sala, Aws Albarghouthi, Caiming Xiong, and Shafiq Joty
    In International Conference on Learning Representations (ICLR) 2026
    Code Paper
  6. NeurIPS
    Beyond Accuracy: Dissecting Mathematical Reasoning for LLMs Under Reinforcement Learning
    Jiayu Wang*, Yifei Ming*, Zixuan Ke, Caiming Xiong, Shafiq Joty, Aws Albarghouthi, and Frederic Sala
    In Advances in Neural Information Processing Systems (NeurIPS) 2025
    Code Paper
  7. EMNLP Oral
    Demystifying Domain-adaptive Post-training for Financial LLMs
    Zixuan Ke, Yifei Ming, Xuan-Phi Nguyen, Caiming Xiong, and Shafiq Joty
    In Empirical Methods in Natural Language Processing (EMNLP) 2025
    Best Paper Nomination
    Code Paper
  8. TMLR
    A Survey of Frontiers in LLM Reasoning: Inference Scaling, Learning to Reason, and Agentic Systems
    Zixuan Ke, Fangkai Jiao, Yifei Ming, Xuan-Phi Nguyen, Austin Xu, Do Xuan Long, Minzhi Li, Chengwei Qin, Peifeng Wang, Silvio Savarese, Caiming Xiong, and Shafiq Joty
    In Transactions on Machine Learning Research (TMLR) 2025
    Paper
Full publication list