What
InfosecTrain, a leading cybersecurity training provider, is hosting a 2-Day LLM Security & Red Teaming Masterclass. This masterclass will be a two-day intensive program focused on the security and adversarial evaluation of modern AI systems, particularly large language models (LLMs).

Day 1 will cover the fundamentals of AI, including machine learning, deep learning, generative, and agentic AI, as well as the full model development lifecycle from data collection to pre-training and fine-tuning. Participants will explore transformer architecture, LLM operation, and practical applications such as chatbots, code generation, and cybersecurity use cases. The session will also highlight frameworks and tools for AI development, including Scikit-learn, TensorFlow, Hugging Face, LangChain, LlamaIndex, OpenAI API, Ollama, and LMStudio, alongside key security considerations during deployment.

Day 2 will focus on AI and LLM red teaming, defining adversarial testing, and exploring attack families like poisoning, injection, evasion, extraction, availability, and supply chain threats. It will dive into LLM-specific risks, including prompt injection, data exfiltration, malicious packages, model poisoning, system prompt leakage, hallucinations, and DoS scenarios. Participants learn hands-on methods using tools like CleverHans, Foolbox, and the Adversarial Robustness Toolbox, providing a practical framework to evaluate, stress-test, and secure AI systems effectively.

When
01-02 November 2025 (Sat-Sun)
7:00 PM - 11:00 PM (IST)

Speaker Lineup
Avnish
7+ Years of Experience
AI Security Expert | Information Security | Cloud Security | Data Security | Consultant & Trainer
Ashish
10+ Years of Experience
AI Red Teaming Expert | Network+ | Security+| Pentest+ | CEH | CND | ECSA | CCNA | ECDE | CPENT | LPT | OSCP

Why Attend
Attending the LLM Security & Red Teaming Masterclass will provide a unique opportunity to bridge the gap between AI development and cybersecurity expertise. Participants will gain a deep understanding of the practical challenges and vulnerabilities associated with large language models, including prompt injection, data exfiltration, and model poisoning. The program will also introduce advanced frameworks and tools for evaluating AI security, allowing attendees to simulate real-world attacks and strengthen defenses. By exploring the full AI lifecycle, from model development to deployment, participants will learn how to identify weak points, implement safeguards, and ensure responsible use of AI technologies. The masterclass will also emphasize hands-on, scenario-based exercises, providing actionable experience in red teaming AI systems. Security professionals, developers, and AI practitioners will acquire the skills to anticipate potential threats, assess risk effectively, and apply mitigation strategies. Ultimately, the program will equip attendees to contribute to safer AI implementations, preparing them for emerging roles in AI security and responsible AI governance.

Agenda (Two Days of Transformative Learning)

DAY 1: Introduction to AI and LLM Security by Avnish (7 PM - 11 PM)
• Demystifying the core concepts and components of an AI system
• Types of AI Systems: Machine Learning, Deep Learning, Generative AI, Agentic AI
• Building and deploying AI - Model Development Lifecycle
• Understanding LLMs: Transformer Architecture, Pre-training and Fine Tuning
• LLM Applications: Chatbots, Code Generation, Cybersecurity Use Cases
• AI and GenAI Frameworks: Scikit-learn, Tensorflow, AutoML, Hugging Face, LangChain, Llamaindex, OpenAI • • API, Ollama, LMStudio
• Security Considerations while Developing and Deploying AI Systems

DAY 2: AI and LLM Red Teaming by Ashish (7 PM - 11 PM)
• Introduction to AI Red Teaming – What is it and why it is needed?
• Attack Families for AI Red Teaming: Poisoning, Injection, Evasion, Extraction, Availability, Supply Chain
• LLM01: Prompt Injection – Direct and Indirect
• LLM02: Sensitive Information Disclosure – Data exfiltration
• LLM03: Supply Chain – Malicious Packages and Models
• LLM04: Data and Model Poisoning – Poisoning datasets and models during training and fine-tuning
• LLM05: Improper Output Handling – Injection via model outputs
• LLM06: Excessive Agency – Agents with dangerous privileges
• LLM07: System Prompt Leakage – Exposing hidden system instructions through crafted queries
• LLM08: Vector and Embedding Weaknesses
• LLM09: Misinformation – Detecting Hallucinations
• LLM10: Unbounded Consumption – Resource abuse and DOS Attacks
• Tools and Frameworks for LLM Red Teaming: Cleverhans, Foolbox, Adversarial Robustness Toolbox

Key Takeaways
• 8 CPE Credits Issued on Completion
• Interactive Red Team Labs
• Attack & Defense Playbooks
• Actionable Security Techniques
• Expert Guidance & Mentorship
• 15+ Cutting-Edge AI Tools

Registration Link: https://www.infosectrain.com/pages/lp/llm-masterclass/

About InfosecTrain
InfosecTrain is a recognized leader in cybersecurity training, focused on enhancing awareness and expertise in data protection, cybersecurity, and compliance. Through expert-led sessions and informative events, InfosecTrain equips professionals and organizations to protect sensitive information and effectively navigate the constantly evolving cybersecurity landscape. With extensive industry knowledge, they are positioned as frontrunners in cybersecurity training and consulting. Additionally, they offer continuous post-training support for future reference, fostering ongoing learning.
To know more about training programs offered by InfosecTrain:
Please write back to [email protected] or call at IND: 1800-843-7890 (Toll-Free) / US: +1 657-221-1127 / UAE: +971 569-908-131