Prompt Injection: The Dark Arts of LLMs (Live Demo & Deep Dive)

Prompt Injection is the most underestimated threat in modern AI systems. If your LLM app can be tricked, hijacked, or manipulated, your entire product stack is at risk. In this session, Nanda Kumar breaks down how attackers exploit LLM trust models and shows live demos of manipulations happening in real time.

This is not a theory talk.
This is what actually breaks AI systems in production.

What you’ll learn in this video:
• How Prompt Injection works and why traditional security models fail
• Real-world attacks that bypass guardrails and safety filters
• How context leakage happens in RAG, Agents, and memory-enabled systems
• Live Demo: Prompt injection, role hijacking, hidden data exposure, and more
• Practical defenses to secure your LLM applications
• Frameworks & tools you can apply immediately to build safer AI systems

Who should watch:
AI Builders, Security Engineers, Developers, Researchers, Tech Founders, and anyone architecting LLM, Agentic, or RAG-powered products.

Key Takeaways:
• Hands-on understanding of Prompt Injection risks
• Security-first mindset for LLM app design
• Demo notebooks and defense playbook
• How to start building AI products that can’t be tricked, manipulated, or weaponized


Our Promise
We don’t teach AI from a textbook — we transfer the expertise and battle-tested experience needed to deploy it in the real world.Our mission is to bridge the gap between AI theory and enterprise transformation, empowering professionals to lead the next wave of intelligent innovation.

🌐 https://saavigen.ai | 🔗 LinkedIn: / nandakumar80 | 🧠 Blog: https://saavigen.ai/article.html


🌐 Learn more:👉 www.saavigenai.com📖 Explore our latest insights on the SaaviGenAI Blog: www.saavigenai.com/blog💼 Follow our updates and discussions on LinkedIn: / saavigenai

Nanda Kumar Kirubakaran
Generative AI Strategist | Founder & CEO, SaaviGen.AI

Nanda Kumar Kirubakaran is a Generative AI strategist specializing in enterprise LLM deployment and security. He founded SaaviGen.AI to help organizations build production-ready GenAI systems that balance innovation with risk management.

Background
With 23+ years in enterprise technology, Nanda held senior leadership positions at Cisco, Hewlett Packard Enterprise (HPE), Aruba, and ChargePoint, where he led large-scale network operations and built high-performing cybersecurity and operations teams. At Cisco, he worked extensively on security product development—including SIEM and NextGen Firewall solutions—helping the organization achieve global compliance standards. His career has spanned cybersecurity consulting, security product development, NOC and SOC operations, and implementation of security compliance programs across global infrastructure.

Current Focus

As founder of SaaviGen.AI, Nanda is committed to guiding professionals and enterprises in their GenAI journey, with a strong emphasis on LLM security. Drawing on his 23+ years in cybersecurity and enterprise technology, he helps organizations implement and safeguard Generative AI initiatives—ensuring solutions are resilient against emerging risks. Nanda engages in industry discussions, leads executive-level sessions, and shares actionable expertise that bridges traditional security rigor with cutting-edge AI advancements. His mission is to empower organizations to unlock AI’s full potential—securely, responsibly, and with lasting impact.

Expertise Areas
AI Security (OWASP LLM Top 10, Prompt Injection Defense)
LLMOps & Production Deployment
Enterprise AI Governance

🌐 [saavigen.ai](https://saavigen.ai) | 🔗 [LinkedIn]( / nandakumar80 )
📍 Location Tags:Bangalore | India | Global AI Security | Enterprise AI Training
#aiethics #genai #aitraining #saavigenai #artificialintelligence #prompting