The tech behind modern AI/ML use cases in Gaming
summary
Explains how real game pipelines ingest event streams with MSK and Kafka, run Flink for parsing/validation/enrichment, persist raw data to S3 and call SageMaker endpoints, and serve features from DynamoDB and Athena for analytics; claims a tested pipeline on m5axlarge instances sustains 60,000 messages/sec with 7 MB average message size; covers RAG with embeddings and vector stores, LongChain and LlamaIndex connectors, continuous fine‑tuning state machines, and practical governance and hallucination risks.
AI-written summary · about
chapters (10)
- 0:00Introduction and setup
- 1:30AI and ML basics
- 9:42Classical ML in gaming
- 16:06Generative AI overview
- 26:03Accelerating game development
- 30:18Enhancing game experiences
- 36:39Intelligent AR and chatbots
- 39:50Gen analytics and dashboards
- 43:41Traditional ML vs Gen AI
- 45:35Q&A session
related talks
26mEngineering Excellence: Platforms and Architectures for Real-Time AI Analytics
42mAgentic AI with AWS Streaming
58mAI-Enhanced Big Data: Integrating Private LLMs and Vector Databases
24mEngaging new and diverse audiences with ML
41mAdding Generative AI to Real Time Streaming Pipelines
18mScaling ML workflows for real-time moderation challenges
16mOperationalizing AI in Finance: Real-Time Analytics
44mBest practices for analytics and generative AI on AWS(ANT302)