Swing
Valuation-led candidates assessed alongside catalysts, momentum, and crowd positioning. Research horizon: days to weeks.
Hwa Gallery & Co. builds AI-native infrastructure for autonomous stock research, decision accountability, and controlled trading workflows. Our first Claude-built agent is already operating in simulated markets; we are now extending that foundation toward broker-connected execution.
*As of the 2026-10-08 Phase 0 report. All fills are simulated paper trades reconciled across the ledger, portfolios, and dated journals.
Our first autonomous research agent was implemented entirely with Claude AI and Claude Code. It investigates U.S. equities, manages three separate paper-trading strategies, records the reasoning behind simulated trading decisions, and reviews outcomes to inform later research. That operational foundation now supports HWA Engine’s next stage of development.
Claude Code was used to build the software. Separately, Claude-based scheduled workflows operate the autonomous research, simulated trades, and retrospective reviews. Both are documented in the project’s own records. All trading is simulated — this is not real-money brokerage execution.
Valuation-led candidates assessed alongside catalysts, momentum, and crowd positioning. Research horizon: days to weeks.
Fair-value discounts, business quality, and durable investment theses. Research horizon: months to years.
Attention-driven research centered on changing social discussion, price confirmation, and defined thesis invalidation.
Modern research is fragmented across market data, narratives, decisions, execution controls, and outcomes. Much of today’s trading software makes execution faster while obscuring the reasoning that precedes it.
HWA Engine is built around the opposite premise: every step from research to action should be recorded, reviewable, and independently constrained.
PHASE 0 STEPS ARE IMPLEMENTED IN SIMULATED RESEARCH · INDEPENDENT CONTROLS ARE PHASE 1 DEVELOPMENT WORK, NOT A LIVE CAPABILITY
With Claude, a founder-operated project can iterate on research, software, and operational workflows at a pace once associated with much larger engineering teams. Phase 0 was built that way, end to end.
That capability doesn’t remove the hard part. Reliable systems still require independent constraints, verification, transparency, and oversight — which is exactly where Phase 1 engineering is focused.
Claude-built AutoStockTracker: simulated portfolios, systematic analysis, trade journaling, and reusable lessons.
Headless Robinhood integration for one strategy, independent risk enforcement, and order-level auditability.
Potential monitoring and administrative UX when operational needs justify it. No committed launch date.
Hwa Gallery & Co. is an independent early-stage venture founded and led by Elaina Lee, focused on AI-native systematic trading infrastructure — currently centered on founder-operated research and private development, not public investment services.
ABOUT THE COMPANYFor research collaboration, technical partnerships, or company inquiries, contact Hwa Gallery & Co. directly.