I need an end-to-end back-office platform that pairs solid database engineering with practical AI so our trading desk can move faster and stay compliant. The core modules should handle transaction processing and record keeping automatically while an AI layer delivers data analysis, detailed historical performance reports, and continuous risk-management insight. Scope of work • Build a scalable database architecture able to ingest, store, and version every trade, position change, and cash movement in near real-time. • Create robust transaction-processing workflows: trade capture, allocations, settlements, and reconciliations must flow straight through with full audit trails. • Design AI models that scan positions and market data to flag risk exposures and exceptions during the trading day, then summarise them in end-of-day reports. • Develop an analytics dashboard that lets us slice historical performance by strategy, symbol, trader, and time period, exporting to PDF/Excel on demand. • Expose clean REST or gRPC APIs so front-office and compliance tools can pull data without manual intervention. Acceptance criteria 1. All trades processed within 2 seconds of receipt, error rate <0.1%. 2. Historical performance report renders under 5 seconds for a one-year data range. 3. Risk alerts generated at least every 60 seconds when limits breached. 4. Database load tests sustain 10k writes/second with zero data loss. 5. Source code, Docker files, and deployment instructions delivered in a Git repo, plus a brief user guide. Preferred stack: Python (FastAPI or similar), PostgreSQL/TimescaleDB, Pandas, scikit-learn or TensorFlow, Docker/Kubernetes, and Grafana or Streamlit for the UI—but I’m open to alternatives if they meet the benchmarks above.