Donato Capitella Courses

Bypassing LLM Guardrails - Anti-Spotlighting and Best of N Attacks

Discover advanced techniques to bypass LLM security measures using anti-spotlighting and best-of-N attacks with hands-on demonstrations in Spikee.

• YouTube
• 47 minutes
• On-Demand
• Free Video

Zero-Click M365 Copilot Exploit EchoLeak - Deep Dive

Uncover the EchoLeak vulnerability in M365 Copilot that enables zero-click data exfiltration through email attacks, exploring RAG system exploitation and defense strategies.

• YouTube
• 38 minutes
• On-Demand
• Free Video

Spikee: A Simple Prompt Injection Kit for Evaluation and Exploitation of LLMs

Explore Spikee, a powerful tool for evaluating and testing LLM security through prompt injection techniques, benchmarking capabilities, and comprehensive guardrail assessment frameworks.

• YouTube
• 29 minutes
• On-Demand
• Free Video

Design Patterns for Securing LLM Agents Against Prompt Injection - Paper Review

Explore six security design patterns to defend LLM agents against prompt injection attacks with practical code examples and implementation guidance.

• YouTube
• 34 minutes
• On-Demand
• Free Video

Finetuning LLMs on Strix Halo - Full, LoRA, and QLoRA on Gemma-3, Qwen-3, and GPT-OSS-20B

Master local LLM finetuning on AMD Strix Halo using full-parameter, LoRA, and QLoRA techniques with Gemma-3, Qwen-3, and GPT-OSS-20B models through hands-on implementation.

• YouTube
• 55 minutes
• On-Demand
• Free Video

Hacking and Securing LLM Applications - Understanding Browser Control Security Risks

Explore security vulnerabilities in LLM applications, focusing on browser and coding agents, with practical demonstrations of prompt injection attacks and essential mitigation strategies.

• YouTube
• 40 minutes
• On-Demand
• Free Video

Multi-Chain Prompt Injection and Jailbreaking in LLM Applications - Security Testing and Defense Strategies

Explore advanced techniques for exploiting vulnerabilities in multi-chain LLM applications, including prompt injection methods, security testing approaches, and essential defensive strategies.

• YouTube
• 47 minutes
• On-Demand
• Free Video

Faithfulness Checks and Hallucination Detection in RAG Pipelines - Lab Tutorial

Master techniques for evaluating and ensuring LLM response accuracy in RAG pipelines using tools like gpt4o-mini, Llama-index, RAGAS, and Azure's Groundedness API for reliable AI applications.

• YouTube
• 38 minutes
• On-Demand
• Free Video

OpenAI CLIP for Zero-Shot Image Classification and Image Similarity - Lab Tutorial

Explore OpenAI CLIP's capabilities for zero-shot image classification and similarity, from local implementation using HuggingFace to deployment on Azure AI Studio with practical examples.

• YouTube
• 22 minutes
• On-Demand
• Free Video

Building LLM Agents - PAL, ReAct, and LangChain Implementation Guide

Master building LLM agents using PAL and ReAct paradigms, implement custom solutions and Langchain framework, and explore various language models while learning function calling and operational loops.

• YouTube
• 1 hour 10 minutes
• On-Demand
• Free Video

Microsoft 365 Copilot Security Vulnerabilities and Attack Vectors

Delve into critical security vulnerabilities in Microsoft 365 Copilot, exploring prompt injection attacks, potential data breaches, and essential mitigation strategies for enterprise LLM applications.

• YouTube
• 21 minutes
• On-Demand
• Free Video

Multi-Modal LLMs for Image, Sound and Video - Episode 6.3

Explore the architecture and implementation of multi-modal LLMs, focusing on Vision Transformers, contrastive learning techniques, and hands-on experience with Google's PaliGemma.

• YouTube
• 24 minutes
• On-Demand
• Free Video

Building LLM Applications Securely - Understanding Risks and Mitigation Strategies

Explore security risks and mitigation strategies for LLM applications and autonomous agents, covering prompt injection vulnerabilities, AI safety misconceptions, and practical defense mechanisms.

• YouTube
• 57 minutes
• On-Demand
• Free Video

Running Gemma 2B and Llama-2 7B with Model Quantization - A Hands-on Lab

Dive into practical implementation of Gemma 2B and Llama-2 7B models, exploring quantization techniques, prompting strategies, and hands-on deployment in Google Colab with free T4 GPU resources.

• YouTube
• 40 minutes
• On-Demand
• Free Video

Fine-tuning DistilBERT for Sentiment Analysis of Restaurant Reviews - Lab 5.3

Master sentiment analysis through hands-on fine-tuning of DistilBERT, using PyTorch and HuggingFace to build a restaurant review classifier with practical implementation steps.

• YouTube
• 21 minutes
• On-Demand
• Free Video