No more selfish silos
UK mobile operator Vodafone is using an AI Booster artificial intelligence system from Google Cloud to teach its systems to get on with themselves and each other and self-diagnose through the machine learning operating service (MLOps). Google Cloud has also claimed it can offer telcos the industry’s ‘cleanest’ service.
Self taught
Vodafone is hoping this single Machine Learning (ML) service can be extended across all its divisions with its learning capacity adapting it to a variety of use cases, according to Vodafone AI Booster product manager Sebastian Mathalikunne, who talked of taking “across markets and vertically from proof of concept to production”.
Compliance
Vodafone claims it has completely automated ML lifecycle compliance activities such as drift and skew detection, articulacy and auditing competence with the use of reusable pipelines, containers and managed services. The approach has also embedded security into the system. Users must complete an online form and in minutes they receive a fully functional AI Booster environment with all the right guardrails, controls and approvals, it claimed in a blog.
Silo silver lining
Vodafone began working with Google Cloud in 2019 to build the Nucleus data system, which is driven by processes such as BigQuery, Dataproc and Cloud Data Fusion. Nucleus is a unified data system which amalgamated all the telco’s other silos of data in its diaspora of customer records, network performance statistics and applications. Previous it had struggled to rationalise incompatible fields and formats hived off into clunky data warehouses. That is ‘in the past’, according to Osman Peermamode, Vodafone’s head of data and analytics who said his department now enjoys the freedom of ‘a single source of truth, making data accessible across our organisation’.
Vantage
For example, Vodafone’s old Vantage installation, which it bought from data warehouse vendor Teradata, will run more smoothly over the Google Cloud Platform, Vodafone claimed. Vodafone first began its transition to its preferred cloud infrastructure in 2019, it claimed in its blog. The new GoogleCloud tool joins a roster including BQML, AutoML and Vertex AI, all installed, integrated discipline and cultivated by Google Cloud systems integrations partner Datatonic.
Datatonic
Datatonic builds reusable MLOps Turbo Templates, a reference implementation of Vertex Pipelines, to expedite the production of the MLOps system for Google Cloud. Jamie Curtis, Datatonic’s Practice Lead for MLOps, explained the secrets behind their economy of effort. “Our team is devoted to solving complex challenges with data and AI. We knew the extent of change Vodafone was embarking on with AI Booster,” said Curtis. Through this open-source codebase, Datatonic’s team created a common standard for installing ML models at scale on Google Cloud. “The benefit to one data scientist alone is significant, so scaling this across hundreds of data scientists can really change the business,” said Curtis.