Beyond the Hype: AI Testing Lessons from the Public Sector
In high-stakes use cases, testing AI systems for reliability and safety is essential. And the stakes don’t get much higher than in the public sector! So, for our August meetup, we thought it would be a good idea to get the public and private sector together.
We have invited a few best-in-class GenAI dev and testing practitioners from the public sector to share their experiences.
👨💻 Testing at scale across multiple use cases/ teams by Benjamin Goh | National AI Group & GovTech
Step into the world of AI Guardian, a groundbreaking SaaS platform to enable GenAI testing across the public sector. Ben will take us through the journey of implementing comprehensive AI testing and share insights on creating standardised safety guardrails across teams through Litmus and Sentinel.
⚖️ Embedding safety and reliability into GenAI for the legal sector by Eric Tan | IMDA Biztech
Discover how GPT Legal is transforming Singapore's legal landscape. Eric will reveal the intricate process of developing an LLM specifically for legal research, highlighting their partnership with Singapore Academy of Law and the implementation of robust safety measures throughout the development lifecycle.
📊 Testing if your RAG application knows its limits by Jessica Foo & Shaun Khoo | GovTech
Get exclusive insights into "KnowOrNot", an innovative approach to detecting hallucinations in RAG applications. Learn how this open-source solution cleverly manipulates context to ensure AI remains grounded in reality – a game-changer for high-stakes applications.
🚀 Why attend?
Get real-world case studies from high-stakes environments
Learn how to design your own testing approach — not just replicate someone else’s
Walk away with practical insights you can apply in your organisation tomorrow
Connect with peers who care about building safe and reliable AI
🔍 Who should attend?
GenAI application developers and testers
Technically-minded business, product and risk stakeholders
This isn't a lecture. It’s a starting point. You’ll leave with the frameworks to test smarter—on your own terms, for your own users.