Sony Interactive Entertainment: Generative and predictive AI on AWS
summary
Sony Interactive builds a governed AI backbone on AWS using DataRobot to run both generative and predictive workloads: automated autopilot tournament model selection, exportable model jars for EMR backtesting and CI/CD deployment, and a Gen AI guard models + human feedback loop for correctness, citations and alerts on hallucinations and LLM cost. The platform gives a 360° console, lineage and hot‑swappable multi-provider LLM playground, enabling reuse across fraud, account security, payments and content search while controlling drift, costs and deployment risk.
AI-written summary · about
chapters (11)
- 0:00Introduction to DataRobot
- 1:03DataRobot AI platform overview
- 3:51Generative AI guard models and feedback
- 6:41Multi-provider LLM playground
- 8:27AI sprawl and technical debt
- 10:15Sony’s ML pipeline journey
- 12:25Generative AI use cases at Sony
- 13:53KPIs and business impact
- 14:37Operationalizing Gen AI models
- 16:02DataRobot in Sony’s pipeline
- 19:59Final takeaways and closing
related talks
16mEnterprise AI Innovation: Gen AI Experimentation
23mGenerative AI use cases
30mGenerative AI in action: from POC to business value
16mEnterprise AI Innovation: GenAI Experimentation
27mBuilding a Future-Ready AI Platform: Pictet AM Journey with AWS
01h 00mSecurely accelerating generative AI innovation
29mIndustry generative AI in action
46mBest Practices for Industrializing Responsible Generative AI