01 Walkthrough
One short demo and two screens from the product.
02 Stack layers
Interface
API
ML & data
03 Architecture
Ingest workers store market and signal data. FastAPI pulls recent features, runs the ensemble, and returns a scored decision.
- React UI for exploring market decisions
- FastAPI + ML ensemble (Random Forest, LightGBM)
- PostgreSQL for market/financial data
04 Engineering challenges
Signal fusion
Market prices and unstructured real-world signals are normalized into one comparable feature space.
Training pipeline
Windowed feature generation and scheduled retraining keep the ensemble current as market regimes shift.
Fast inference
Model loading, caching, and lean feature queries keep decision endpoints responsive.
Explainability
Tree-based models expose feature importance, so a decision can always be traced to its drivers.
05 Trade-offs & decisions
Time-series joins across instruments, signals, and predictions are fundamentally relational.
Async IO and typed contracts fit ML serving; OpenAPI docs come free.
On tabular financial data, gradient boosting wins on accuracy per unit of training cost, and it stays explainable.