From smart antennas to auto-configuring core and RAN elements, AI-driven automation is poised to redefine how telecom networks are deployed, maintained, and optimized. Increasingly, engineers are stepping back from manual parameter tuning to focus on data, decision-making, and intelligent orchestration.
Why the Shift Is Happening
- RAN complexity explosion: The advent of 5G New Radio (NR) with its myriad band combinations, slices, and cloud-native architectures has made manual tuning unsustainable. Operators must now embrace AI/ML to handle everything from link adaptation to secondary-carrier prediction.
- Autonomous core networks: GenAI enabled processing of traffic, user behavior, and voice fraud detection is now viable in 5G/6G cores .
- Capital and Opex efficiency: Data-driven RAN automation, like Ericsson’s intent-based frameworks, reduces manual O&M and spectrum/energy waste.
What Engineers Will Do Less… and More
What’ll shrink:
- Parameter tuning for RRM, power control, slicing
- Manual core config updates
- Reactive troubleshooting
What’ll grow:
- Data engineering: building pipelines, lakes, and ETL with high-quality, trustworthy input for AI systems.
- Feature engineering and model ops: identifying meaningful KPIs, integrating KPI feedback into closed loop control, and deploying scalable model lifecycles (MLOps & AIOps).
- AI software development: crafting smart agents and intelligent controllers for open/RAN and edge environments .
- Explainability & trust: especially in O-RAN, engineers must make AI decisions transparent and auditable.
- Cross-domain orchestration: integrating RAN, core, transport, and edge intelligence for holistic AIOps.
Skills Gap & Upskilling Roadmap
- A recent study revealed 33% of telecom engineering roles lack 5G/Open RAN, data, and AI/ML competency.
- Data engineering is emerging as a foundational discipline within AI engineering handling everything from ingestion to pruning.
- GenAI is moving from lab experiments to scaled core use cases, and demands strong data pipelines and model oversight .
What This Means for Telecom Engineers
- Invest in next-gen education: Learn Python, SQL/NoSQL, ETL tools, and ML fundamentals. Familiarize yourself with MLOps and AIOps practices.
- Tangible AI experience: Join internal RAN/CORE AI pilots or collaborate on open/O-RAN AI projects.
- Build domain-specific pipelines: For example, gather and label real-time KPI streams to support anomaly detection (e.g. LSTM in Cloud-native RAN analytics) .
- Champion transparency: Deploy explainable AI strategies in your RICs, easing trust hurdles .
- Embrace intent-based design: Shift to high-level intent inputs with AI-driven orchestration, instead of manual RAN config.
Telecom networks are evolving from hardware-centric systems into data-rich, intelligent platforms powered by AI. That means both core and RAN engineers are well-positioned but only if they evolve alongside.
The future is not about pushing bits from point A to B; it’s about engineering the data and intelligence that automatically handles those bits at scale.
