Collector
Four Hubs, sixteen Thrusts and current graduate application windows verified.
P0 Institution / 2026-08-04
HKUST
Guangzhou.
Institutional diligence across Agent Infrastructure, Data+AI systems, AI4Science and AI Quant / Financial Technology.
P0Build three supervisor tracksNo outreach before evidence review
Executive judgment
One institution. Three credible routes.
The official Hub/Thrust structure makes Agent Infrastructure, Data+AI and FinTech legitimate cross-disciplinary paths. The opportunity is high, but a generic university application would be weak. PPHD recommends separate Data+Agent, Agent/AI4Science and AI Quant evidence packages.
Analyzer
Yuyu Luo, Chao Zhang and Sijia Chen form the first high-information shortlist.
Future
Cross-disciplinary option value is strong; unfocused breadth is the main risk.
Writer
Write one proposal per track. Never combine Data, Science and Quant into one vague pitch.
Interviewer
Defend novelty, non-Agent baseline, evaluator validity and implemented evidence.
Campus Helper
Plan for Nansha, English research communication and cross-Hub research rhythm.
Future Development
Strong routes into Agent Infra, Data+AI, AI4Science and financial ML research.
Relationship
Reproduce, benchmark, obtain feedback, then request Human-approved outreach.
Paper Collector
Freeze three recent papers for every shortlisted supervisor before ranking.
Paper Writer
Use a governed Agent fixture to connect PAT with Data or Quant evaluation.
Three-track strategy
Do not apply to an institution. Apply to a research problem.
Data+Agent Systems -> Yuyu Luo and adjacent DSA faculty Agent / AI4Science -> Menglin Yang, Zhijiang Guo, Chengwei Qin, Jintai Chen AI Quant / FinTech -> Chao Zhang, Sijia Chen and adjacent FinTech faculty
Each track needs its own task, baseline, experiment, contribution and target-specific research proposal.
First diligence shortlist
Highest-information conversations.
Yuyu Luo
Foundation Agents, AI for databases, data-centric AI and public engineering systems.
Chao Zhang
ML+Finance, LLMs, asset pricing, volatility, graph structure and price impact.
Sijia Chen
Large reasoning models, multi-Agent collaboration, simulation and risk control.
Menglin Yang
RAG, Agent Memory, reasoning, retrieval and AI4Science.
Zhijiang Guo
Rigorous reasoning, robust knowledge integration, reliability and scalability.
Chengwei Qin
LLM reasoning, Agent systems, multimodality and efficient learning.
Jintai Chen
Trustworthy multimodal and multi-Agent systems for healthcare and discovery.
Xuming Hu
Trustworthy, efficient and vertical large language models.
Highest-IG action
Reproduce before outreach.
Run one Data+Agent or Agent+Finance baseline, freeze its Benchmark Card, and use the result to choose the first supervisor-specific conversation. No bulk email and no admission claim.