AstraQuant is an event-driven algorithmic trading engine designed to detect and execute high-probability continuation moves following institutional momentum.
The system implements a strict behavioral model:
Momentum Burst → Market Acceptance → Controlled Continuation
Rather than predicting price, AstraQuant reacts to verified order-flow behavior while enforcing hard risk constraints.
AstraQuant is built on three principles:
- No prediction. Only trade confirmed behavior.
- Risk first. Execution is always subordinate to session and drawdown guards.
- Deterministic infrastructure. Every action is logged, reproducible, and auditable.
The platform is deployed as a containerized microservices stack:
| Service | Role |
|---|---|
engine |
Core decision engine & trade orchestration |
ws |
Market data ingestion and event streaming |
ai |
Statistical validation & adaptive filters |
risk |
Hard risk gates and session governance |
api |
Execution interface & broker integration |
db |
Trade + telemetry persistence |
redis |
Event buffering / coordination |
nginx |
Controlled external exposure |
All services run via Docker Compose for deterministic deployment.
- Session drawdown guard
- Loss-streak pause logic
- Execution enable/disable switches
- Environment-level trade blocking
- Full execution logging for auditability
AstraQuant is designed for VPS-based 24/7 execution.
/opt/astraquant
├── docker-compose.yml
├── engine/
├── risk/
├── Strategy services...
This layout separates infrastructure from user environments and mirrors production Linux service conventions.
Live research and controlled execution environment.
This repository tracks infrastructure and strategy evolution — runtime data, logs, and secrets are intentionally excluded.
This project is a research system. It is not financial advice and is operated with strict internal risk controls.