Create Docker containerization for reproducible deployment #4
Description
Summary
Package the backtesting engine into a Docker container for reproducible runs across environments and easy deployment to cloud infrastructure.
Technical Approach
- Write a multi-stage Dockerfile: first stage installs build dependencies and compiles native extensions (XGBoost, LightGBM, TensorFlow), second stage copies only the runtime artifacts to minimize image size.
- Create a docker-compose.yml with services for the engine itself plus optional Redis (for result caching) and a lightweight dashboard.
- Add a Makefile with common targets: `make build`, `make test`, `make run STRATEGY=momentum_rsi` for quick iteration.
- Pin all Python dependency versions in a lock file to guarantee bitwise-reproducible backtest results across machines.
- Document the container workflow in README with examples for running on AWS EC2 / GCP Compute.