About
Building the infrastructure for autonomous capital allocation.
Founder & Principal Architect
Shawn Wang
Shawn spent the first decade of his career building distributed systems at scale. At AWS, he shipped production infrastructure across Amazon Connect and Aurora Global Database — the same database engine backing some of the world's highest-throughput workloads. He then founded IntelliLead, shipping an AI-powered SaaS product from concept to paying users in two weeks, winning first place at Dragon's Chamber Pitching Competition. He co-founded VEST AI for investment research, served as CTO at EarnAha, and worked as Principal Solutions Architect at Nextlink, where he led Cathay Group's (Taiwan's largest financial institution) digital modernization and taught cloud-native ML to PhD candidates at National Taiwan University College of Medicine.
In late 2025, Shawn turned his full attention to the intersection of AI and capital markets. Six months of dedicated research and system development produced: a production-grade algorithmic trading engine, ML prediction pipelines with rigorous walk-forward validation, source-level competitive analysis of a dozen AI agent systems, and a comprehensive architecture for autonomous multi-agent trading. Every hypothesis tested. Every decision rationale documented. Every failed idea killed with data, not opinion.
Opal Research is the synthesis of a decade of building production systems and a deep conviction that autonomous AI agents are the next computational primitive for markets — the way quantitative models were the last one.
Why autonomous agents. Why now.
Frontier models are improving fast. Every capability gain translates directly into better research, better debate, better risk assessment — if the architecture is built to absorb it. The fund that already has an autonomous colony architecture compounds every model improvement. The fund that routes everything through human approval hits the same ceiling regardless of model quality.
We're building the architecture ahead of the models. When GPT-6 or Claude 5 arrives, Opal's colony doesn't need a rewrite — it gets smarter overnight. That's the structural advantage of building AI-native rather than AI-assisted.