Telco Cloud 5G Artificial Intelligence Beginner
This project turns the shared-lab patterns into a 5G service-assurance workflow. Participants build three structured agents — alarm triage, KPI / RCA, multimodal field evidence — and then a fourth completely freehand agent that they design from scratch. The freehand lab is the hackathon-vibe finisher the customer asked for.
.svg){kind=small}
About The Lab
Prerequisites
Audiences
Lab Architecture
The Lab architecture seamlessly integrates Claude Code as the operator interface, a multi-model layer (Claude Haiku for all agents, Pixtral-large for vision, faster-whisper for audio), and MCP connectors to live 5G observability (Grafana MCP with the free5gc-5g-assurance dashboard) to create a comprehensive environment for AI-driven 5G operations, which includes three specialized labs, Alarm Triage, KPI/RCA, and Field Evidence — that allow the user to test different end-to-end NOC use cases, from classifying alarm clusters and detecting KPI anomalies with root-cause hypotheses to analyzing multimodal field inputs (images and voice notes), producing structured JSON outputs (triage_output, rca_output, field_report) ready for downstream action.
Why this Lab ?
This project contain 5G service-assurance workflow. Participants build three structured agents -- alarm triage, KPI / RCA, multimodal field evidence -- and then a fourth completely freehand agent that they design from scratch. The freehand lab is the hackathon-vibe finisher the customer asked for.
Lab Objectives
- Build three cooperating 5G agents following structured guidance, then a fourth freehand agent.
- Use real model routing for alarm triage, KPI reasoning, and field photo and audio interpretation.
- Combine static incident artifacts with live mock MCP data covering alarms, KPIs, ownership, and tickets.
- Practice human-in-the-loop approval gates so AI suggestions are always reviewed before execution.
- Finish the project with one freehand agent that reflects something the learner actually wants in their day job.
