Title

AWS re:Invent 2022 - Goldman Sachs: Accelerating time to value in data analytics (FSI201)

Summary

Gerard Cowburn, Senior Solutions Architect at AWS, introduces the session on accelerating time to value in data analytics.
Ram Rajamoni, VP and Tech Fellow in Data Engineering, and Francesco Pontrandolfo, Product Manager for Goldman Sachs Financial Cloud, join the session.
The session covers how Goldman Sachs blends diverse hybrid data sources to inform investment hypotheses and deliver insights quickly.
An overview of Goldman Sachs and its data-driven investment process is provided.
The architecture of Goldman Sachs Financial Cloud is discussed, focusing on high-speed, low-latency, real-time analytics for financial market data.
Key data sourcing integrations and the use of the open-source Legend platform for data modeling and wrangling are examined.
The session concludes with insights and lessons learned to help others accelerate their data analytics journeys.

Goldman Sachs heavily invests in technology, with a significant portion of its workforce being engineers.
Time and speed are critical factors for competitiveness in the financial industry.
The Goldman Sachs Financial Cloud is a modular set of services addressing data curation, management, and analytics.
Data curation includes access to GS proprietary data and third-party data sets with an added curation layer.
Compute instances can be spun up on the GS Financial Cloud, allowing for the ingestion and enrichment of data with a consistent data model.
Data analysis tools include REST endpoints, a Python SDK called GSQuant, and a data visualization tool called Portal Pro.
The architecture of GS Financial Cloud on AWS includes ECS, DynamoDB, ElastiCache, OpenSearch, and a custom time series database optimized for AWS.
Real-time market data is streamed into the platform using a bespoke API and a solution called Electron.
Challenges in financial data include its evolving nature, the breadth and structure of data sources, and the need for real-time data streaming.
Cloud scalability, managed serverless infrastructure, and early engagement in risk-managed data transfer are key enablers for overcoming these challenges.