🌟 Hello Azure Qatar Community! 🌟
We’re excited to bring you a session that takes on one of the most important questions in AI today: How do we make large language models (LLMs) give answers we can trust?

💡 𝗠𝗮𝗸𝗶𝗻𝗴 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗠𝗼𝗱𝗲𝗹𝘀 𝗠𝗼𝗿𝗲 𝗥𝗲𝗹𝗶𝗮𝗯𝗹𝗲: 𝗛𝗼𝘄 𝗧𝗼 𝗗𝗲𝘁𝗲𝗰𝘁 & 𝗥𝗲𝗱𝘂𝗰𝗲 𝗛𝗮𝗹𝗹𝘂𝗰𝗶𝗻𝗮𝘁𝗶𝗼𝗻𝘀
Large Language Models (LLMs) are reshaping industries - but they can still hallucinate: confidently providing answers that sound correct but are factually wrong.
👀 What you’ll learn in this session:
✅ A crash course on what hallucinations in AI are, why they happen, and why they’re such a challenge for real-world use
✅ How clustering techniques can help detect hallucinations by spotting patterns of inconsistent or contradictory answers when the model is uncertain
✅ How Sparse Autoencoders (SAEs) can enrich user queries in a way that guides models toward more accurate, grounded responses
✅ Examples of how these techniques could help get more reliable results from AI systems
🎁 BONUS: Microsoft Build 2025 AI Highlights

After exploring the core topic, we’ll give you a quick recap of key Azure AI announcements from Build, including: Azure AI Foundry, Agents Service, Azure AI Search Multimodal RAG and more!
🎙 Speaker:
Samir Abdaljalil
Samir is a PhD candidate and Graduate Research Assistant at Texas A&M University, where his work focuses on making large language models more trustworthy and reliable. His research spans natural language processing, machine learning, and AI safety, with a particular focus on detecting and reducing hallucinations in AI-generated text.
Samir holds a bachelor’s degree in Information Systems from Carnegie Mellon University and has rich background in Natural Language Processing.
📅 Date: Tuesday 17th June 2025
⏰ Time: 6:30 PM - 8:00 PM
📍 Location: Microsoft, 27th Floor Burj Al Fardan, Lusail
📝 Register now: https://events.azureqatar.org/er9w8/ — seats are limited!
📺 Can’t attend in person? Watch the livestream on YouTube: https://www.youtube.com/@AzureQatarCommunity/streams
✨ Whether you’re a developer, researcher, or just curious about AI’s future, this is your chance to learn how to spot and fix AI hallucinations — and get a glimpse of the latest tools coming to Azure AI.