Case Study: APMIC achieves faster, lower-cost AI model training with Google Cloud Platform

APMIC Delivering hyper realistic AI assistants for a wide-range of scenarios

APMIC, a Taiwanese natural language processing company, needed more powerful compute resources and a more efficient way to train and run its machine learning models for chatbots, text analysis, and other enterprise AI services. To address these scaling and availability challenges, APMIC migrated its workloads to Google Cloud Platform, using services such as Compute Engine and Vertex AI.

With Google Cloud Platform, APMIC streamlined its ML pipeline, improved training convergence, and cut environment setup time from one week to one hour. The company also reduced operational workloads by 50% with GKE Autopilot and lowered ML operational costs by 50%, while eliminating database failures with Cloud SQL and improving service availability for its 3.35 million users across 720 enterprise customers.

APMIC

Jerry Wu

Founder and CEO