Jay Shah
Cyber Security and DevSecOps Professional
Toronto, Canada
Actions
Jay is a Cyber Security and DevSecOps Professional. He is a Technology & Educational consultant and has 7+ years of Industry experience & has received his Master's degree from Southern Methodist University, Texas, USA. He is Academic Liaison for Canada DevOps Community of Practice, Vice-Chair IEEE YP and worked in companies like Jio, Supreme Lending, Cyber Group & EY. He has various Cloud & Cyber Security Certifications. He is Ambassador for Continuous Delivery Foundation & The DevOps Institute
Area of Expertise
Open-Source Generative AI: Exploring IP Ownership, Security Risks, Legal Regulations & Frameworks
Open source Generative AI models generally offer varied benefits such as lower initial and maintenance costs, broader community support, faster response and speed to market, better code transparency, flexibility etc. as compared to closed source GenAI models. But at the same time, use of open source GenAI may impose varied risks such as possibility of trade secret & data leakage, output bias and hallucinations for users. Further, developers could also face copyright issues in case copyrighted data is used without permission and licensing risks for output code created etc. Multiple governments and regulatory bodies across the globe are undertaking initiatives to draft policies, standards for copyright protection, risk mitigation, data and national security and legal compliance etc. to ensure safe, secure and ethical use of Generative AI.
We will cover open & closed source Generative AI Models, Intellectual property data ownerships, risks, regulatory and legal compliance. The talk will also enlighten various key security & practical considerations in choosing open-source AI models for business workflows along with industry- led lessons learnt, best-practices and frameworks used.
Mastering Multi-Modal AI: From RAG Fundamentals to Industry Applications
In this session will demystify Retrieval Augmented Generation (RAG) models, which represent a pivotal advancement in AI by combining retrieval mechanisms with generative models to produce more accurate and contextually relevant outputs. Starting with the foundational concepts of RAG, including the roles of the retriever and generator, we'll explore how these models outperform traditional approaches. We'll give a walkthrough on the process of training RAG models & introduce the concept of Multi-Modal RAG, which integrates both text and image data, opening new avenues for AI applications and use-cases across various sectors and industries and end-to-end development to deployment process and public Cloud platform Generative AI service for Azure and Oracle.
We will also look at some of the Industries Best Practices, Methodology and also talk on some of the Industry wide Cyber Security Frameworks including NIST, ISO and Model Risk Management.
Navigating the Future: Responsible AI Through Ethics, Governance,CyberSecurity Frameworks & Controls
The adoption and activation of Large Language Models (LLMs) introduces various cyber risks, including biases, data privacy, trust and ethics concerns, and cyber threats throughout the AI model lifecycle. Enhancing AI frameworks with dynamic risk assessment models, robust data security privacy measures, and adaptive learning algorithms can address these challenges. Bias in AI algorithms is a significant ethical consideration in the use of AI in cybersecurity, and it is essential to take steps to identify and mitigate potential biases to ensure fair and responsible use of AI. This includes using diverse and representative training data, technical solutions such as adversarial training and fairness constraints, and governance structures, policies, and procedures.
In this panel, our speakers will talk and discuss on AI principles, AI lifecycle management, model risk management framework, confidence in AI along with NIST, ISO cyber security controls and risk mitigation techniques.
Please note that Sessionize is not responsible for the accuracy or validity of the data provided by speakers. If you suspect this profile to be fake or spam, please let us know.
Jump to top