Inspiration
Trading tools often examine models, trades, positions, and P&L separately. Doctor Quant was inspired by the opportunity to connect them in one investigation workflow.
What it does
Doctor Quant is a read-only flight recorder for trading algorithms. It records what an algorithm knew, why it decided to buy, sell, or hold, what was executed, the resulting position, and its P&L.
It also provides a common framework for connecting different APIs, data providers, models, and broker records.
How we built it
We researched quantitative trading, AI observability, data provenance, and trade auditing. We built the evidence system and server in Python, with a JavaScript, HTML, and CSS dashboard.
Challenges we ran into
The main challenges were preventing future information from affecting historical explanations, supporting strategies with different indicators, and separating intended trades from actual executions.
Accomplishments that we’re proud of
Doctor Quant can dynamically display different strategy inputs, replay decisions using only information available at the time, and identify missing or contradictory evidence instead of inventing explanations.
What we learned
We learned that understanding a trading decision requires more than a chart. It requires reliable timestamps, decision records, model versions, executions, positions, and accounting data.
What’s next for Doctor Quant
Next, we want to add real broker and market-data integrations, improve execution and P&L reconciliation, and make Doctor Quant easy to embed into existing trading systems.
Log in or sign up for Devpost to join the conversation.