Full ML platform buildout on GCP — enabling AI-powered search, personalisation, and recommendations for a major B2B distributor with a legacy technology base.
A major European B2B electronics distributor operating at scale with a legacy search infrastructure and no ML platform. Product discovery was entirely keyword-driven, with no personalisation or ranking intelligence. The organisation lacked any ML infrastructure for experimentation or production deployment.
Legacy search was delivering poor product discovery results at scale — customers were not finding the right products, leading to drop-offs and lost revenue. There was no experimentation framework to test improvements, no feature store for ML models, and no path to deploying AI-powered features in production. The technical foundation needed to be built from scratch before any AI capability could be delivered.
We built the enterprise AI platform on GCP with full MLOps practices — experiment tracking, feature store, model registry, and production serving infrastructure. On top of this foundation, we deployed three AI-powered systems in sequence: AI-powered search with semantic understanding and behavioural ranking, personalisation at the individual customer level, and a product recommendation engine. Each system was deployed with A/B testing capability and production monitoring from day one.
Let's talk about what's achievable for your specific context.
Start the Conversation →