This massive global scale brings significant latency, data, and infrastructure limitations and operational constraints, and after managing its own technology stack using on-premises data centers, Booking.com turned to AWS to build its end-to-end data strategy. Using AWS, Booking.com has seen a three-fold increase in model training jobs, a two-fold decrease in failed jobs, and a five-fold decrease in training time.
Booking.com is also in the early phases of exploring how Amazon Bedrock and Amazon Titan might further amplify its work in machine learning and artificial intelligence. Booking.com recently introduced an AI Trip Planner, which uses Amazon Sagemaker to enhance retrieval augmented generation to provide highly personalized customer accommodation recommendations. Booking.com has also adopted the open-source Llama 2 model from AWS to build an intent detection system with seven billion parameters. With the support of AWS, Booking.com is not only…
