Chih-Hsuan (Bella) Yang
Agentic AI Research and Applied Machine Learning
bellayang@anl.gov (mailto:bellayang@anl.gov)  |  chihhsuan-yang.github.io (https://chihhsuan-yang.github.io/)
GitHub (https://github.com/ChihHsuan-Yang)  |  Hugging Face (https://huggingface.co/ChihHsuan-Yang)  |  Google Scholar (https://scholar.google.com/citations?user=s7RDC3YAAAAJ)  |  AgentsSci (https://huggingface.co/AgentsSci)  |  LinkedIn (https://www.linkedin.com/in/chih-hsuan-yang-isu/)
Argonne postdoctoral researcher in LLM agents, multi-agent systems, and agentic evaluation, with hands-on preference-model training, DPO/PPO, and AI for Science experience.

Technical Skills
Agents: evaluation harnesses, protocol orchestration, trace diagnostics, memory ablations.
Post-training: preference/reward modeling, RLHF, DPO, PPO, scientist feedback.
Implementation: Python, PyTorch, scikit-learn, NumPy/pandas, Hugging Face, Git.

Research and Industry Experience

Postdoctoral Researcher, Argonne National Laboratory | Jul 2025 - Present
- Build reproducible agent evaluation harnesses with matched protocol runs, structured traces, and token/call accounting.
- Analyze reviewer-to-solver critique uptake, answer repair, and task-conditioned memory utility through controlled comparisons.
- Evaluate cost-aware protocol routing with supervised baselines, held-out tests, confidence calibration, and quality-cost metrics.
- Release research code and evaluation data; participate in AuroraGPT's Post-Training Working Group and mentor four students.

Visiting Scientist, Corteva Agriscience | Sep 2024 - Jan 2025
- Trained preference models from scratch using experimental outcomes to ground rewards; incorporated scientist preferences when judgments were uncertain.
- Adapted multiple language models with Direct Preference Optimization (DPO) and Proximal Policy Optimization (PPO).

AI Resident, X, Google LLC | Sep - Dec 2023
- Developed stakeholder-based multi-agent LLM frameworks; studied information flow, coordination, communication, and bidding.

AI Graduate Researcher, Iowa State University | Jan 2020 - Jun 2025
- Co-led BioTrove; researched molecular language models and multi-fidelity materials design.

Selected Publications

Precise but Uncoupled (https://chihhsuan-yang.github.io/NeurIPS26_Precise-but-Uncoupled/) | 2026
First author. NeurIPS Main Track, accepted.

Cost-Aware Protocol Routing (https://aclanthology.org/2026.findings-emnlp.1275/) | 2026
First author. Findings of EMNLP, published.

BioTrove (https://baskargroup.github.io/BioTrove/) | 2024
First author. NeurIPS Datasets and Benchmarks, Spotlight.

Education
Ph.D., Mechanical Engineering, Iowa State University | 2025
Co-major in Computer Engineering; minor in Applied Mathematics.
M.S., Computer Science, Iowa State University | 2025
B.S., Bioenvironmental Systems Engineering, National Taiwan University | 2019