CV

AgentSuite: Toward More Reliable Agent Evaluation with a Component-Based Benchmark Auditing Pipeline

H. Suh*, S. Lee*, B. Ji*, R. Khare, B. Khan, H. Kim, Tianyi Zhang, et al. (* equal contribution)

ICML 2026Paper

Causal Brain Connectivity: Integrating Granger Directed Graphs in fMRI Analysis

Tianyi Zhang, K. Han, J. Nie, et al.

AIME 2025Paper

Brain Network State Transformer: Leveraging State Functional Connectivity

J. Nie, K. Han, Tianyi Zhang, et al.

IEEE EMBC 2025 · OralPaper

Distributed Training Chaos Lab

An open-source self-check harness for distributed training failures — runs real PyTorch training jobs, injects controlled stressors, and turns DCGM/NVML, NCCL, XID, thermal, power, PCIe/NVLink, memory-pressure, and correctness signals into reproducible run reports for GPU-cluster preflight and triage.

coact

A governance substrate for two coding agents in one repo — lets Claude Code and Codex work in the same project safely via filesystem coordination (capability policies, locks, a task board, worktree isolation, an append-only journal). Single dependency-free Go binary.

Wealth Projection Simulator

A bilingual, 100% client-side wealth-projection simulator (v1.0) — projects net worth year by year (nominal vs. real) from careers with equity/RSU, home/mortgage, and market assumptions; adds Monte Carlo (p10/p50/p90), stress tests, FIRE, and goal-seek ("$X by age Y"). Installable PWA; no backend or tracking, open-source (MIT).

NVIDIA NeMo RL

Contributor to NVIDIA's NeMo RL — a scalable library for reinforcement learning and post-training of large language models.