Transforming pharmaceutical R&D through industrialized machine learning
summary
AstraZeneca builds an AWS-based ML platform that runs hundreds of projects with hundreds of active users, cutting environment setup from months to days. A single-tenant, multi-account AI Bench 2.0 gives each project three accounts (service automation, dev/pre-prod, prod), audit logs and CI/CD, and qualifies as GxP compliant. Tooling includes SageMaker + Lambda and Dataiku. A deployed NLP pipeline for literature surveillance outperforms human review, increases throughput, and improves as models retrain.
AI-written summary · about
chapters (19)
- 0:00Introduction to AI in R&D
- 0:37Data and AI Vision
- 2:16Data Governance and FAIR
- 3:00Data Science Platform
- 4:44R&D Challenges
- 5:24AI Applications in R&D
- 7:13Strategic Approach
- 8:48ML Operations Hurdles
- 9:23Key Operational Challenges
- 12:58Goal of Democratization
- 14:04Progress and Solutions
- 14:48Internalizing R&D Data
- 16:11Partnership with AWS
- 16:38Evolution of AI Bench
- 17:46AI Bench 1.0 Details
- 19:30Lessons Learned
- 19:59AI Bench 2.0 Features
- 22:34Literature Surveillance Case
- 25:46Future Outlook
related talks
20mFrom promise to impact with gen AI in healthcare & life sciences
32mAutomating Production Production Level Machine Learning on AWS
36mNew trends in data modernization: Catalysts for AI/ML analytics on AWS
04mCustomer Spotlight
44mA generative AI–enabled enterprise: Transformative AI/ML on AWS
45mStart your AI/ML Projects Right
30mML Platform for Insurance Conglomerate
26mAccelerating innovation in the healthcare and life sciences industry