Collector
Identity and supervisor role verified. Current funded opening remains unverified.
P0 Target / 2026-08-04
Professor Chao Huang
& HKUDS.
Public-source diligence on Data Intelligence Lab@HKU, its Agent-native research trajectory, Pengyi fit, application strategy and Greater Bay Area comparison set.
P0Prepare target-specific pitchVerify current opening before outreach
Executive judgment
Application-driven Agent Infrastructure.
The strongest intersection is research on how retrieval, memory, tools, planning and evaluation change reliability and cost in high-constraint Quant workflows. HKU sources identify Professor Huang as an Assistant Professor and PhD supervisor leading Data Intelligence Lab@HKU, with work spanning LLMs, AI agents and graph machine learning.
Analyzer
Very high Agent-system fit; experimental isolation and publication evidence are the main gaps.
Future
Trajectory moves from graph/recommendation into RAG, runtimes, ecosystems and domain Agents.
Writer
Pitch one controlled research question, not the number of Agent Teams built.
Interviewer
Expect attacks on novelty, baselines, leakage, variance and generality beyond Quant.
Campus Helper
Plan HKU research rhythm and Shenzhen-Hong Kong mobility only after a verified path exists.
Future Development
Strong transfer to Agent Infra, open-source AI systems and AI-native Quant research.
Relationship
Reproduce, publish evidence, obtain technical feedback, then request approved outreach.
Paper Collector
Map retrieval, runtime, research/coding and domain-application project clusters.
Paper Writer
Ablate capabilities against success, leakage, cost and human-intervention metrics.
Candidate RP
Verifiable Quant research as an Agent benchmark.
Deterministic baseline + Base LLM Agent + PAT-enabled variants + Capability ablations + Success / leakage / cost / intervention evaluation
The contribution must be a controlled result about which capability improves which failure mode. “An AI Quant platform” is not a sufficient novelty claim.
Similar-group map
HKU first. Greater Bay Area next.
Data Intelligence Lab
Chao Huang: LLMs, agents, graph learning and open-source systems.
IoT Lab
Edith Ngai: Edge General Intelligence and LLM multi-agent collaboration.
AI-Agents in Business
Multi-agent platforms and domain-specific business Agents.
SAIL / InfoBodied
Multimodal, spatial and embodied Agent research.
Xuming Hu / Jian Guo
Trustworthy LLM systems and AI-for-Quant research; supervision routes require verification.
Zhongxiang Dai / Mengnan Du
LLM Agents, routing, optimization, reasoning and trustworthy AI.
AI Theory & Systems
AI infrastructure, chips, automated benchmarks and Agent optimization.
Language Model & HCI
Agentic intelligence and human-AI society; identify exact supervisor and route.
Highest-IG Action
Freeze the RP outline and Benchmark Card by August 6.
Then verify the current recruitment route and opening before asking Pengyi to approve outreach. No autonomous email or admission claim.