How Tripadvisor Delivers Real-Time Personalization at Scale with ML | HackerNoon
Briefly

Tripadvisor uses advanced machine learning models and ScyllaDB on AWS to deliver real-time personalized recommendations to users. The platform operates at an impressive scale, managing over 2 billion requests daily from 25 to 50 million users. It ensures millisecond-latency for personalized interactions, serving 400 million unique visitors every month. At peak times, Tripadvisor can reach around 425,000 operations per second with impressive read and write latencies between 1-3 milliseconds. The system is supported by a team of over 2,800 professionals and generates more than $1.8 billion in revenue annually.
Tripadvisor's personalization engine uses machine learning models and ScyllaDB on AWS to process over 2 billion requests daily, providing real-time recommendations to users.
Each click on Tripadvisor's platform is processed in milliseconds, ensuring that personalized recommendations are delivered at a scale of 400 million unique visitors each month.
At peak traffic, Tripadvisor achieves approximately 425,000 operations per second on ScyllaDB, with read and write latencies kept at around 1-3 milliseconds.
With a team of over 2,800, Tripadvisor generates more than $1.8 billion in revenue, showcasing its significant impact on global travel and hospitality.
Read at Hackernoon
[
|
]