An Open Source Architecture for Modernizing Market Risk Management
summary
Proposes an open-source, real-time event-driven backbone that moves risk from overnight batch to sub-minute evaluation using VAR and Monte Carlo (10k–100k sims) and even per-second margin checks; ties Kubernetes/OpenShift, Kafka/Pulsar, Airflow, Jupyter/Spark and GPU acceleration to S3/Ceph caching and Hazelcast/Redis for in-memory compute; claims orchestration cuts friction so quants push models into production faster, surfaces that 80% data movement is the bottleneck, and can auto-hedge on breach via automated escalation.
AI-written summary · about
chapters (10)
- 0:00Modernizing market risk
- 1:22Risk management fundamentals
- 3:51Drivers for modernization
- 6:18Modern risk analytics features
- 9:47Hybrid cloud strategy
- 11:52Real-time risk orchestration
- 14:18Data movement challenges
- 17:30Event-driven risk calculations
- 20:17Model development acceleration
- 22:03Summary and next steps
related talks
01h 06mA modern, scalable risk system architecture
36mA Data Science Approach to Systemic Risk
15mArchitecting Scalable Platforms for Financial Wellness
35mScaling Climate Data for FinTech with an Open Source Data Mesh
13mBuild Real-Time Analytics Platforms for Financial Crime Detection
47mCreating a Contemporary Risk Management System Using Python
32mBuilding Real-time Systems with Open Source Technologies
15mPlatform Engineering for FinTech