Workflows

From idea to reviewable research run

DQT orchestrates internal quant workflows, not data subscriptions or profit guarantees.

Workflow 01

Idea to backtest

Turn a research idea into an executable experiment while preserving parameters, sample windows, and computation records.

Research Agent structures the hypothesis, universe, and statistical relationship to test.

Creates candidate features and marks required internal data plus missing checks.

Backtest Agent calls deterministic engines and records parameters plus OOS windows.

Outputs risk, significance, sensitivity, and suggested next steps.

Workflow 02

Data is not an external product. It is the internal foundation that lets agents understand research context, provenance, quality, and versioning.

Connect internal research data, derived market tables, and approved sources.

Align fields, time, asset mapping, and versions.

Detect gaps, anomalies, latency, and coverage changes.

Provide source links and quality summaries to agent runs.

Workflow 03

Turn operating state, risk events, and approvals into an internal cadence instead of scattering them across messages, notebooks, and spreadsheets.

Collect risk, task, data, and backtest operating state.

Trigger notifications for anomalies, threshold breaks, and failed jobs.

Connect events to sources, strategy versions, and recent changes.

Generate daily notes, weekly briefs, and owner review items.