Learn Series: Do You Trust Your AI?

--> 0 % Overview ​Large language models sound confident even when they get the facts wrong. If you’ve ever caught yourself believing an AI output a little too quickly, this evening is for you.

​Join the Beyond the Vibes Community at SGInnovate for an interactive session breaking down what happens beneath the surface of confident model outputs, how trust is evaluated, and when you should double-check your tools.

​ The Hallucination Trap: Why LLMs generate persuasive prose even when factually incorrect.

​ Evaluating AI Agents: How developers and researchers measure reliability, drift, and calibration.

​ Trust Frameworks: Practical methods to determine when to automate and when to require human review.

​ Interactive Q&A: An open floor to discuss failure modes and edge cases from daily workflows.

​ 7:00 PM – 7:10 PM: Opening Remarks by SGInnovate & Beyond the Vibes Community

​ 8:00 PM – 8:40 PM: Session 2: LLM-as-a-Judge Is Lying To You

​ Focus: Cloud Data Architectures, Generative AI & Developer Tools

​ Su Myat Noe — AI Research Scientist, National Institute of Informatics, Tokyo

Wei has been lucky enough to work with some of the largest technology companies in the world on all things data. He has a deep background and passion in data engineering.

Su Myat Noe is a AI Research Scientist at the National Institute of Informatics (NII) in Tokyo. She focuses on multimodal AI safety and evaluation. She has presented her research at leading international conferences, including the ICML 2026 AI and many more. She spend most days thinking about questions like: What happens when AI makes mistakes? How do we measure safety and reliability? Can these systems behave responsibly in the real world?

Terms of Use • Privacy & Cookies Policy • Report Vulnerability