Introduction to Large Language Models (LLM) in 5G: Enhancing Intelligence

Course details

Telecom networks generate vast amounts of data—from network logs to customer interactions—and AI is essential for extracting value from it. This course provides a practical foundation for applying large language models (LLMs) and generative AI to real-world telecom challenges. Explore core LLM and transformer concepts, then build custom named entity recognition models for network log analysis, develop text and intent classification systems for customer support automation, and fine-tune pretrained models for telecom-specific use cases. The course covers the complete pipeline—from data preparation and annotation to model evaluation—with a focus on business impact, not just accuracy metrics. By the end, you’ll be equipped with the skills and strategic framework you need to take AI projects from pilot to production, integrating AI-driven insights into CRM systems and operational workflows.

Learning objectives
Explain the foundational concepts of generative AI (GenAI), including the role of transformers, the structure and capabilities of large language models (LLMs), and how tokens function as fundamental units for building contextual semantics.
Differentiate between token classification and text classification techniques, and evaluate their respective applications within a telecom decision-making framework, including named entity recognition (NER) and intent classification.
Design and simulate the process of building a custom NER model for telecom network logs, including data preparation, tokenization, annotation, fine-tuning of pretrained models, and business-impact-focused evaluation.
Analyze how to extract domain-specific information by building customer-specific and compliance-focused NER models, and construct a telecom entity taxonomy that bridges generic language models to industry-specific terminology.
Assess the integration of intent classification and sentiment analysis models into business logic.
Formulate strategies for converting model outputs into actionable insights within CRM systems and customer support workflows.
Develop a strategic telecom AI road map that outlines the journey from pilot projects to production deployment, incorporating best practices for model evaluation, accuracy improvement, and business value realization.

Concepts

Introduction

Introduction

Introduction to Large Language Models (LLMs)

Overview of generative AI concepts
Working principles of generative AI
Role of transformers in GenAI
Large language models - Structure and capabilities
Industrial transformation through language models
Tokens as the fundamental units of large language models
From tokens to contextual semantics

Core Concept - Token Classification

What is token classification From theory to telecom
Named entity recognition (NER) and beyond
Visualizing token classification - From chaos to structure

Core Concept - Text Classification

What is text classification Automating categorization at scale
Few-shot learning and intent classification
Why LLMs have transformed data and performance
Preview - Building a telecom-specific intent classifier

Building a Custom NER Model for Network Logs

Building a custom NER model for network logs
Architecture - Classification layer for token labeling
Data preparation and tokenization pipeline for telecom logs
Data preparation and tokenization pipeline - Simulation
The annotation process - Creating telecom training data
Fine-tuning a pretrained model for custom entity recognition - Simulation
Evaluating your NER model - Beyond accuracy to business impact

Extracting Information - Building Models

Building models for customer-specific entity recognition
Compliance-focused NER model for telecom - Simulation
Building a telecom entity taxonomy - From generic to domain specific

Automating Customer Support with Intent Classification

From model output to business logic - Creating actionable intent rules
Evaluating model performance and improving accuracy

Sentiment Analysis for Customer Experience

Combining intent and sentiment - The complete telecom customer picture
Sentiment analysis for telecom customer experience
From analysis to action - Integrating insights into CRM
Building your telecom AI road map - From pilot to production

Conclusion

Conclusion