Leveraging AI and LLMs: From Technical Foundations to Business Operations
Course taught by Daniele Teti
Duration: 1 day
Price: € 490.00 VAT excluded
Next course on February 26th, 2026
Prerequisites:
Knowledge of Delphi or Python. Basic understanding of how an API with an HTTP/S endpoint works.
Description
The goal of the course is to provide practical skills to select, implement, and integrate local LLMs in business contexts, highlighting their advantages, limitations, and requirements. The course focuses on local Large Language Models (LLMs) for enterprise applications. It is designed to be hands-on, avoiding theoretical or mathematical deep dives, and aims to help participants understand which systems to use, their benefits, limitations, and the necessary requirements for effectively implementing local LLMs in a business setting. The structure is divided into modules with hands-on activities to ensure practical learning.
Provided Materials
Documentation and Guides
- Complete course slides in PDF format
- Technical documentation for implementation and deployment
- Reference guides for APIs and configurations
- Best practices for integration in business environments
Source Code and Projects
- Complete Delphi projects (VCL and FMX) for all examples shown
- Ollama/PostgreSQL/DelphiMVCFramework integration libraries
- SQL scripts for vector database setup
- Modelfiles configured for different use cases
Templates and Resources
- System prompt templates for specific domains
- Example datasets for testing and validation
- Docker configurations for deployment
- Checklist for production implementation
Support and Guarantees
- Complete materials: Access to all source code and documentation used in the course
- Certificate: Certificate of participation with training hours
The course provides all the necessary materials to understand and implement AI solutions with Delphi: complete documentation, working source code, configurable templates, and technical support to guide implementation in your business projects.
Contents
MODULE 1 – AI Scenario and Business Opportunities (45 minutes)
- AI in Italian companies: current situation
- Analysis of ISTAT data and adoption trends
- Main barriers: skills, costs, implementation
- ROI evaluation in corporate AI projects
- LLMs for business applications
- Automation of document processes
- Technical assistance and customer service
- Automatic analysis and classification of content
- Practical example: Implementation of a corporate chatbot
MODULE 2 – Limitations of LLMs and Solutions (1 hour)
- The 4 fundamental limitations of LLMs
- Context size and token management
- Knowledge cut-off date and obsolete information
Hallucinations: recognizing and preventing them - Lack of access to external data
- Strategies to overcome the limitations
- Context optimization techniques
- Information validation methods
MODULE 3 – Implementation with Ollama (1.5 hours)
- Ollama setup and configuration
- Installation and model management
- Optimization for corporate hardware
- Configuration for multi-user environments
- Integration with Delphi applications
Ollama- REST API: robust implementation
- Error and timeout management in production contexts
- Integration patterns for existing applications
- Practical exercise:
- Integration of AI functionalities in an existing VCL application
- Implementation of a contextual help system
- Performance testing and optimization
MODULE 4 – Personalization and System Prompt (1 hour)
- Designing corporate system prompts
- Methodologies for defining specific behaviors
- Adaptation to company domains and procedures
- Techniques for quality control of responses
- Creation of specialized models
- Using Modelfile for customization
- Templates for different sectors (legal, financial, technical)
- Validation and testing of personalized models
- Practical lab: Development of a specialized assistant for your domain
MODULE 5 – Data Analysis and Chain of Thought (1 hour)
- Processing structured corporate data
- Techniques for analyzing CSV, Excel, and JSON files
- Patterns for extracting information from documents
- Validation and quality control of results
- Chain of Thought for decision-making processes
- Implementation of multi-step reasoning
- Application to validation and control processes
- Traceability of automated decisions
- Case study: Automatic sales data analysis system
MODULE 6 – Vector Databases and RAG System (1 hour)
- Implementation of vector databases
- PostgreSQL configuration with PGVector
- Document embedding and indexing techniques
- Chunking strategies for corporate documents
- Development of RAG (Retrieval Augmented Generation) systems
- Architecture and main components
- Implementation with Delphi and DelphiMVCFramework
- Integration with existing document management systems
- Exercise: Implementation of a corporate knowledge base with semantic search
MODULE 7 – Function Calling and Systems Integration (45 minutes)
- Function calling for functional extensions
- Implementation of calls to external functions
- Integration with existing corporate APIs (database, CRM, ERP)
- Patterns for complex workflow automation
- Model Context Protocol (MCP)
- Implementation of integration standards
- Use with DelphiMVCFramework for enterprise services
- Deployment and maintenance strategies
- Final project: Development of an AI assistant integrated with corporate systems
