Intelligent Notification Systems at Scale
summary
System designs for billions of notifications daily with target 1ms response latency using microservices, event sourcing/CQRS, gRPC and Kafka + Redis (plus RabbitMQ/Pulsar for routing) with multilevel caching and chaos testing. ML uses XGBoost/LightGBM and temporal models, feature store Feast, MLflow/Airflow pipelines, Azure Data Explorer/BigQuery for feedback; predictive delivery timing and adaptive rate limiting drive results: open rates +35%, engagement +45%, mute/uninstall −18%, and an e‑commerce case shows 42% cart recovery and 37% higher CTR.
AI-written summary · about
chapters (15)
- 0:00Introduction and Speaker Background
- 0:26Current Notification Landscape
- 1:16Microservice Architecture for Notifications
- 2:45Machine Learning in Notification Systems
- 3:43Message Brokers Comparison
- 4:58System Resilience Strategies
- 6:32Machine Learning Integration
- 7:23Smart Filtering with ML
- 8:59Targeted Content and Personalization
- 10:57Continuous Improvement and Feedback Loops
- 15:28Predictive Delivery Timing
- 20:35Multi-Level Caching Strategies
- 21:35Case Study: E-commerce Platform
- 23:03Implementation Roadmap
- 24:21Conclusion and Contact Information
related talks
14mPerformance Optimization in High-Volume Transaction Systems
34mBuilding a scalable multichannel notification system
32mLiveMLP: ML Platform for Assisting Contact Center Agents in Real-Time
20mBuilding Intelligent Retention Engines: Real-Time Churn Detection
23mML Platform Challenges & Applications at Freshworks
16mProduction MLOps for Spam Detection: Real-Time Multi-Modal AI
39mBuilding an event-driven notification system in Go and RabbitMQ
08mAI-Driven Innovation for Enhanced Food Delivery Operations