MortadhaMannai/Devfest-tripoli-libya-2025---Bulding-Agentic-AI-workflows-with-xAI-on-Google-vertex-AI-
Créateur:
Mor
Hôte:
DevFest Tripoli Libya 2025 -Building Agentic AI Workflows with xAI on Vertex AI From Code to Action - Overview This repository contains the complete demo, code, and resources used in my DevFest Tripoli Libya 2025 session. The session explores how to design and deploy autonomous AI agents @vertexAI
# **DevFest Tripoli Libya 2025 -Building Agentic AI Workflows with xAI on Vertex AI From Code to Action**
**Author : Eng. Manai Mohamed Mortadha**
---
## **Overview**
This repository contains the complete demo, code, and resources used in my DevFest Tripoli Libya 2025 session. The session explores how to design and deploy autonomous AI agents capable of perceiving information, reasoning with ML models, acting in real time, explaining their decisions, and running at cloud scale using Google Cloud Vertex AI.
The included demo is a local version of a production grade agentic workflow built to reflect real Vertex AI pipelines, decision orchestrators, and xAI tools.
---
# **Session Summary**
Agentic AI refers to AI systems that can:
• Gather information
• Interpret or predict
• Decide autonomously
• Take real actions
• Provide transparent explanations
• Maintain audit trails
This talk demonstrates how to build such systems end to end using Google Cloud Vertex AI, including:
• Decision pipelines
• Model serving
• xAI interpretability
• RAG based retrieval
• Automated agent actions
You will learn both the architectural patterns and the exact code required to deploy your own agent workflows.
---
# **Session Agenda**
## **1. Introduction**
• Why agentic AI matters today
• Where autonomous systems are used
• What makes AI trustworthy
## **2. Understanding Agentic Systems**
• Key components
• Decision flow patterns
• Perception to action loop
• Where xAI fits
## **3. Vertex AI Deep Dive**
• Workbench for development
• Pipelines for orchestration
• Endpoints for real time inference
• Built in explainability
• Scalable serving options
## **4. Building Autonomous Pipelines**
• Defining agent tasks
• Connecting data sources
• Handling realtime triggers
• Automating decisions
## **5. Integrating Explainable AI**
• Feature importance
• SHAP style reasoning
• Attributions
• Transparent inference
## **6. Using Vertex AI Models**
• Predictive logic
• RAG for m …