Executive Summary
Developed as part of the Deutsche Bank Hackathon 2026, Eurorisk Deep Mind addresses the European banking sector's 2026–2028 supervisory priorities: navigating geopolitical volatility, cyber resilience, and strict compliance with the Digital Operational Resilience Act (DORA).
Core Architectural Pillars
- Autonomous Multi-Agent AI: Deploys multi-agent AI systems utilizing Vector Retrieval-Augmented Generation (RAG) to dynamically monitor complex credit risks, Commercial Real Estate (CRE) loan exposures, and sovereign bond mark-to-market losses.
- DORA & Operational Resilience: Automates real-time tracking of critical ICT vendor dependencies, third-party concentration risks, and regulatory compliance standards across European banking jurisdictions.
- Proactive Shock Simulations: Replaces static quarterly reporting with daily, automated macro-economic and geopolitical shock simulations to proactively forecast bank liquidity and capital adequacy ratios.
- Cloud-Native Scalability: Engineered with enterprise-grade Python backend architectures, containerized microservices, and interactive executive dashboards for Chief Risk Officers (CROs).
Technology Stack
- Backend: Python, FastAPI, Pydantic
- AI / LLMs: LangGraph, Vector RAG, Ollama / Vertex AI
- Cloud Infrastructure: Google Cloud Platform (GCP), Cloud Run, Docker
- Regulatory Frameworks: DORA, EBA Supervisory Guidelines