# Yizirui Fang

Canonical page: https://yiziruifang.com/
Website owner: Yizirui Fang

Yizirui (Easey) Fang: Amazon SDE and applied ML researcher working on reliable LLM agents, embodied AI, conformal prediction, and human-AI collaboration. Explore projects, papers, and public evidence.




## About Me

I am an Amazon SDE and applied ML researcher focused on reliable AI agents, embodied/human-centered AI, and uncertainty-aware decision systems. I am interested in work that combines production-grade agentic systems, agentic RL, rigorous model evaluation, and trustworthy agents.

My strongest through-line is turning ambiguous model behavior into measurable scientific problems: code-generation agents and tool-use workflows in production settings, spoken instruction following for embodied agents, learning-to-defer systems for human-AI collaboration, and conformal prediction under data and distribution shifts.

I publish research as **Yizirui Fang** and also go by **Easey Fang**. Find my work on [GitHub](https://github.com/yizirui), [Google Scholar](https://scholar.google.com/citations?user=C9Wb_2cAAAAJ&hl=en), and [LinkedIn](https://www.linkedin.com/in/easey-f/).


## Research interests

- LLM Agents and Tool Use

- Post-training and Evaluation

- Human-centered Machine Learning

- Embodied AI

- Uncertainty Quantification

- Robust Decision Making


## Projects

- [Learning to Defer with an Uncertain Rejector via Conformal Prediction](https://yiziruifang.com/project/uncertain-defer/): TMLR 2026 paper on uncertainty-aware learning to defer, using conformal prediction to make human-AI routing safer under rejector uncertainty and distribution shift.

- [Pragmatic Embodied Spoken Instruction Following in Human-Robot Collaboration with Theory of Mind](https://yiziruifang.com/project/siftom/): ICRA 2026 paper on spoken instruction following for embodied agents using theory-of-mind inference over speech, human perception, and robot goals.

- [ARCHED: Human-Centered AI-Assisted Instructional Design](https://yiziruifang.com/project/arched/): A human-centered framework for transparent, responsible, and collaborative AI-assisted instructional design.

- [Investigating Data Usage for Inductive Conformal Predictors](https://yiziruifang.com/project/conformal-predictors/): Studying how data allocation choices affect inductive conformal prediction, calibration behavior, and uncertainty guarantees.

## Publications

- [Learning to Defer with an Uncertain Rejector via Conformal Prediction](https://yiziruifang.com/publication/fang-2026-learning-tmlr/) — Transactions on Machine Learning Research (TMLR), 2026.

- [Pragmatic Embodied Spoken Instruction Following in Human-Robot Collaboration with Theory of Mind](https://yiziruifang.com/publication/ying-2024-siftom/) — IEEE International Conference on Robotics and Automation (ICRA 2026), 2026.

- [ARCHED: A Human-Centered Framework for Transparent, Responsible, and Collaborative AI-Assisted Instructional Design](https://yiziruifang.com/publication/li-2024-arched/) — Proceedings of the Innovation and Responsibility in AI-Supported Education Workshop (iRAISE 2025), PMLR 273:94–104, 2025.

- [Learning to Defer with an Uncertain Rejector via Conformal Prediction](https://yiziruifang.com/publication/fang-2024-learning-workshop/) — NeurIPS 2024 Workshop on Bayesian Decision-making and Uncertainty, 2024.

- [Investigating Data Usage for Inductive Conformal Predictors](https://yiziruifang.com/publication/fang-2024-investigating/) — arXiv preprint arXiv:2406.12262, 2024.



---

[All papers](https://yiziruifang.com/publication/) · [Research](https://yiziruifang.com/research/)