Faculty Profile: Prof. Nan Jiang — UIUC RL Theory
Faculty Profile: Prof. Nan Jiang — UIUC RL Theory
Position: Assistant Professor Institution: UIUC, Dept. of Computer Science Website: https://nanjiang.cs.illinois.edu/ Report date: 2026-06-12
Research Focus
Theoretical reinforcement learning: sample complexity, function approximation, offline RL and evaluation, partial observability (POMDPs), game-theoretic RLHF and LLM alignment theory.
Academic Profile
- NSF CAREER Award (2022)
- Sloan Research Fellow (2024)
- Google Research Scholar (2024)
- PhD: University of Michigan
Research Direction on LLMs
- RLHF and LLM alignment: game-theoretic approaches to alignment, theoretical analysis of preference-based RL
- Offline RL evaluation: applicable to evaluating RLHF pipelines
- Focus is on mathematical foundations, not engineering LLM systems
Fit with Weijia Zhang
| Dimension | Assessment |
|---|---|
| RL theory / foundations | ✅ World-class |
| RLHF / alignment theory | ✅ Theoretical analysis |
| Applied RL post-training (SFT/GRPO) | ❌ Engineering, not his focus |
| LLM agents / agentic AI | ❌ Not applied work |
| GUI / VLM / multimodal | ❌ Not his area |
Verdict
Pure RL theorist with excellent recognition. Highly relevant if Weijia wants deep theoretical RL foundations (e.g., for a theory-leaning PhD thesis). Not a match for applied agentic AI engineering work. Could serve as a committee member if thesis includes theoretical RL components.