Investors face fragmented evidence
Chinese and English sources, filings, charts, notes and backtesting tools can sit in separate places, making it difficult to maintain one research trail across markets.
Proof case · co-founder / CTO
Mini Bloomberg Terminal
A Chinese-first AI financial research and quant workspace showing how forward-deployed AI practice can transfer into another high-trust domain where sources, boundaries and auditability matter.
Case summary
Chinese-first multi-agent workspace
The public portfolio describes a research OS for global Chinese investors across A-share, Hong Kong, U.S. and Taiwan markets, combining daily briefs, filings intelligence, company research, risk alerts, ETF research and quant experiments.
Governed research infrastructure
The value is not an investment promise. It is a CTO-level example of carrying source discipline, auditability, quality evaluation and clear product boundaries into another high-trust domain.
How the research workflow is organised
Bring recurring tasks together
A single workspace can present daily briefs, filings interpretation, company research, risk alerts, ETF research, quant experimentation and data-map checks as connected research activities instead of isolated tools.
Separate work by responsibility
The architecture described in the portfolio separates data ingestion, filings analysis, company research, risk monitoring, quant lab work, editorial review and compliance checking into specialist agent workflows.
Keep the evidence trail visible
Source grading, citation discipline and audit trails make the assistant easier to inspect. Model choice, data strategy, security and quality evaluation remain part of the CTO responsibility rather than being hidden behind a single chat interface.
Evidence and limits
Alvin Liao
Author and practitioner: Alvin Liao, described in the public portfolio as co-founder and CTO for Hong Kong Press Agency AI Trading Limited.
20 August 2026
This public proof-case summary was updated on 20 August 2026. The prototype link is public; this page does not infer performance or investment outcomes.
Research assistant boundary
The public evidence describes a prototype and product direction. It explicitly positions the platform as a research assistant with citations, source grading, audit trails and regional compliance rules.
No investment advice
The product does not provide buy/sell calls, target prices or personalised portfolio advice. Financial decisions remain with the user and appropriate professional judgement.
Related reading and service
For the wider operating principle, read How can EHS teams use AI without weakening governance? For an organisational approach to readiness and controls, see the Corporate AI Safety Diagnostic.