Research before automation
Hypotheses are tested and documented before anything is automated. Every call records an explicit expectation and an invalidation condition — before the outcome is known.
Hwa Gallery & Co. is a small, independent, founder-led technology venture developing AI-native trading infrastructure. We believe ambitious technical systems can emerge from focused experimentation, disciplined engineering, and the capabilities of modern AI.
The company’s first system — a Claude-built autonomous research and paper-trading agent — already operates in simulated markets. The work now underway extends that foundation toward controlled, broker-connected execution.
Hwa is derived from the founder’s own name. The “Gallery” is a metaphor rather than a corporate history: a gallery is a space where work is placed in the open to be examined — and that is how we treat research. Theses, decisions, reviews, and lessons are hung where they can be revisited, compared, and judged on their own terms.
The name is a commitment to that kind of exploration: ideas made visible, examined honestly, and kept accountable to their outcomes.
Before Hwa, Elaina built and operates The Trading Gallery — an independent retail and e-commerce business spanning multi-location operations, inventory strategy, internal systems, and process and workflow automation — the unglamorous work of scaling the day-to-day systems a real business runs on.
Building real businesses shaped the founder’s interest in software that makes complex processes repeatable, observable, and more accessible to small teams. Hwa Gallery & Co. applies that operator’s discipline to a harder domain: autonomous research and controlled trading infrastructure.
Small is not a limitation here; it is a working method. Practical discipline, direct feedback loops, and founder involvement in every technical decision mean nothing ships on momentum alone. Experiments are documented, assumptions are surfaced, and building comes before scaling.
A founder-led project can move quickly, but it also has to be rigorous about what is actually working, where assumptions remain, and what risks cannot be delegated to an AI system. We prefer documented experiments and incremental progress over inflated claims about scale.
The future of automation should be more accountable, not less.
AI has narrowed the distance between a technical concept and a testable prototype. Claude enabled both the development and the operation of Phase 0: it built the software, and its scheduled workflows run the autonomous research, simulated decisions, and retrospective reviews.
That capability doesn’t remove the hard part. Sound validation, accountable operations, and clear boundaries remain essential — and the most important engineering work begins at the point where a proposed AI action might affect a real account.
Hypotheses are tested and documented before anything is automated. Every call records an explicit expectation and an invalidation condition — before the outcome is known.
Sources, decisions, and retrospective lessons are recorded as they happen, so every claim about the past can be traced to what was written down at the time.
Risk permissions and oversight stay explicit. Automation is earned through testing, bounded actions, and monitoring — never assumed.
A long-term builder philosophy: make capable, accountable systems accessible to small teams. This is a direction of work, not an existing commercial service.
A Claude-built agent researching equities and paper trading three simulated books, with journals, retrospective reviews, and persistent lessons.
Secure broker connectivity with independent safeguards and appropriate authorization — under development, not a live capability.
An optional monitoring and administration interface, if operational needs justify it. No committed date.
Inquiries about the company, research, and technical collaboration are welcome at elaina@hwagallery.com. We are not offering investment advice, managing customer assets, or making a consumer trading product available.
For research collaboration, technical partnerships, or company inquiries, contact Hwa Gallery & Co. directly.