Workflows

The core pipeline an operator walks every time they ship parameters into production: calibrate, validate, configure, paper, real, analyze.

Steps6

Modules Touched4

Money PathCalib → Real

Read SideAnalytics

Step 1 of 6

Calibrate

Produce optimized parameters from historical data

calibration

What you do

Pick a symbol + calibration type (Activity / Microprice / GLFT / Volatility / Liquidity). Pick a date range. Launch — the backend dispatches an isolated docker run and tracks it as an MLflow experiment.

What it produces

MLflow experiment ID, run artifacts under MLflow's artifact store

Surface

/calibration

API calls

startCalibrationRun()fetchCalibrationRuns()

Full pipeline at a glance

Calibration produces optimized parameters MLops promotes them to production Trading wizard loads them Paper trade validates the behaviour Go-live ships real-money execution Analytics monitors P&L, fills, and latency end-to-end.