Machine Learning Skills Gap in the Telecom Industry
Balancing operating efficiencies and top-line growth is a key challenge for any Mobile Network Operator (MNO) executive. Automation initiatives by MNOs have yet to deliver large efficiency gains as promised. Profit squeeze in the Telecom industry continues to drive consolidation and pursuit for efficiencies while preparing to spend about $0.5T of CapEx to upgrade network infrastructure to deliver 5G services.
MNOs compete broadly on two fronts, horizontally with other MNOs for market share and vertically with Over-the-Top(OTT) players for value share beyond connectivity. With 5G services rollout planned to start 2020, the competition for value share is going to intensify.
OTT players have invested significantly in Data Science teams and continue to put Machine Learning (ML) at the core of product design. For instance, Netflix uses ML not just to make recommendations for movies and TV series, Netflix analyzes viewing patterns and scene replays to determine which scenes were of higher impact to their customer experience. They use this intelligence in content creation, curation, and individual recommendations. If MNOs want to compete in this vertical beyond just connectivity, ML skills need to be at the core of new service development. Same is true for Autonomous Vehicles, Healthcare and Retail verticals where MNOs want to claim higher value share.
From an efficiency perspective, MNOs have driven some efficiency gains from labor rate arbitrage and process automation over the last decade. Machine Intelligence is the third frontier for improving efficiencies.
The case for increasing ML skills to turbocharge process automation and create higher value services for MNOs is clear. However, the skill deficit for machine learning is critically large and growing especially in the face of large network transformation trends for 5G readiness.
If you are in network engineering, you need to know the potential of various supervised learning models in optimal capacity planning, optimal demand response, and agile network design. If you are in network operations, you need to know how to apply deep neural nets to predict faults, execute predictive maintenance & diagnose network faults in increasingly complex multivendor networks. If you are a product manager, you need to know how to monetize network-based intelligence to create new revenue streams for your employer. Same is true for product managers for vendors that serve MNOs to provide them with products that are enabled with ML services to serve the needs of your customers.
There are many ways the Telecom industry can build Data Science capabilities:
Hire experienced Data Scientists - Overall Data Science skill shortage makes hiring and retaining a significant challenge. Furthermore, they lack the domain expertise and context of various business use cases.
Depend on existing software and hardware vendors- Traditional Telco vendors are in the process of building their own Data Science teams and are unable to serve their customers in a meaningful way.
Workforce skills upgrade- Internal training organizations are largely dependant on various MOOCs in the training marketplace. However, MOOCs are not designed with a focus on serving use cases in the Telecom industry and lack context in practical implementations for various Telecom Networks.
DaVinci Networks is bringing to market the industry first ML training program for Network Engineers, Architects and Product Managers. Our ML training programs focus on ML technologies relevant to the type of data encountered in Telecom networks. We incorporated practical considerations on ingesting network data such as telemetry & syslog to build machine intelligence. Furthermore, we connected these practical considerations to the evolution of network automation platforms that can take action based on intelligence.
DaVinci Networks has partnered with Aarna Networks to offer a combined training course on ML technology and Open Network Automation Platform (ONAP), which is a Linux Foundation project for real-time, policy-driven orchestration and automation of physical and virtual network functions. ONAP offers built-in capabilities to support data collection, data prep, analytics and closed-loop automation using machine intelligence. Check out our training offerings and see which one fits your organization needs the best.