Technical Workshop: (Hands On) Finetuning GenAI for Hacking and Defending
[Book My Seat]

Trainer: Jitendra Chauhan & Abhisek Datta & Vignesh Chandrasekaran

Jitendra Chauhan (IIT Kgp Alumni)

Creator of Bad Llama, How to Turn Good Llama into a Toxic Llama | Chief Scientist/Co Founder, Detoxio.ai

Jitendra is a seasoned professional in Product & Strategy with over 18 years of experience in the tech industry. An alumnus of IIT Kharagpur,he has been a founding member of two startups.. With three patents to his name, Jitendra's expertise spans AI/ML, Data Engineering, and cybersecurity, where he has built SIEM solutions in the cloud and
developed products focused on Penetration Testing, EASM, and Red Teaming.

Abhishek Datta (Black Hat Speaker)

Creator of Vet, Open Source Supply Chain Tool (Product Hunt 37th Rank) | CTO/Co-Founder, Detoxio.ai

Abhishek is a seasoned expert in Research & Technology, boasting over 18 years of experience and co-founding two ventures. His significant contributions to cybersecurity include identifying over 15 CVEs (Common Vulnerabilities and Exposures). As a speaker and trainer, he shares his deep knowledge in exploitation, malware analysis, GenAI security, and AI supply chain attacks, showcasing his comprehensive expertise and dedication to advancing the security landscape.

Vignesh Chandrasekaran (Balccon Speaker)

Vignesh is an Experienced Information Security researcher with a demonstrated history of working in the information security industry. Skilled in pen testing, Vulnerability Assessment, Red teaming, purple teaming, blue teaming, Threat intelligence, and information technology professional.

>>Block My Seat For Workshop Session  

Workshop Summary: 

Workshop Duration: 4 Hours, 1/2 Day
Date: 30th May, Thursday, 2024 

Description:

This workshop is designed for a technical audience including Appsec Pentesters, Security Engineers, Security Architects, and AI/ML practitioners, Software Developers who have started exploring GenAI. 

Workshop Agenda:

0. Introduction to LLMs: Running an LLM for Fun and Profit

• Overview of Large Language Models (LLMs) and their potential for various applications.
• [Hands-On] Step-by-step guide to setting up and running an open-source LLM.
• [Hands-On] Introduction to interacting with LLMs for specific use cases.

1. Understanding LLMs History, Evolution and Internals

• Deep dive into the architecture of LLMs, focusing on Transformers and Encoders.
• A brief history of the evolution of LLMs, highlighting the advancements from second-generation deep learning models.
• Exploring the distinctions between LLMs and earlier deep learning frameworks.

2. Penetration Testing and Red Teaming LLMs

• 2a. Identifying LLM Application Vulnerabilities
  • Discussion on the various classes of vulnerabilities specific to LLM applications.
• 2b. Hands-on - Testing LLMs and General AI Applications
  • Practical exercises on testing LLMs and GenAI applications for selected classes
    of vulnerabilities.
  • Open Source Tools to Red Team a GenAI Application
• 2c. AI Supply Chain Risks
  • Exploration of risks in the AI supply chain, including model poisoning and
    mitigation strategies.
  • Hands-On - Creating and Detecting Model Poining Attacks

3. Security Monitoring and Guardrails for LLMs

• [Hands on ] Using a Open Source / Community Verison Solution to Develop Security
  Monitoring
• Strategies for implementing security monitoring of LLMs to detect and respond to
  potential threats.
• Guidelines for establishing guardrails to prevent specific types of attacks on LLMs and
  GenAI applications.

>>Block My Seat For Workshop Session 

Candidate Requirements:

• The course is beginner friendly
• Some python programming language and machine learning experienec will help

You need to bring:

Hardware: 

• A laptop with internet access
• A Kaggle account
• A Hugging Face account
• To get Maximum Benefit, Python Programming and Machine Learning Experience Will Help

Who Should Attend ? 

• This workshop is designed for a technical audience including Appsec Pentesters, Security Engineers, Security Architects, and AI/ML practitioners, Software Developers who have started
  exploring GenAI.

Set Expectations:

• Hands on session

>>Block My Seat For Workshop Session