AI Practice and Data Engineering Practice, GovTech

In this post, we’ll discuss why LLM guardrails are essential, and how we think about designing and implementing them at GovTech.

In today’s data-driven world, businesses need insights but are inundated with vast amounts of data. Technologies like Agentic AI and…

Laying the Groundwork for Scalable, Custom AI: Cloud Model Hosting and the Rise of Agent-Based Intelligence

Tools, MCP, and why it matters

Systematically discover challenging edge cases using LLMs

Abstract

Abstract

⚠️ Warning ⚠️ This article may contain offensive material, and is presented here solely for educational and research purposes only.

Deploying models on live production systems is challenging and improving latency is one of the biggest challenges; we can also reduce costs

5 Lessons from My GovTech Internship

How we developed a fast, lightweight zero-shot classifiers for user prompt’s relevance to the system prompt.

With LLMs, we can train our own customised guardrail classifiers, especially in pre-production when we have little data.

In this post, we provide seven plus one technical tips on how to get started with LLM guardrails.

In Part 2 of the Multimodal AI series, find out more about how we used Multimodal AI for cataloguing of digitised physical assets

How much is too much?

“Technology,” “Artificial Intelligence,” “Data-driven decisions” — these are buzzwords nearly everyone has heard of at some point. They…

Hey everyone! I’m Shae-Anne, a final year Data Science and Economics student at the National University of Singapore (NUS). During my…

Read more about the AI Practice’s approach towards Multimodal AI, our explorations, and use cases for the Singapore government.

Find out how we developed a moderation classifier for toxic and unsafe Singlish texts that outperforms commercial alternatives

The spread of misinformation and strong need for efficient content management were challenges that emerged in today’s digital landscape.

Introduction

Getting more out of your model

From the GovTech AI CapDev Team:

How GovTech and Resaro worked together to conduct a fairness assessment on LLM-generated draft testimonials

Shoutout to our co-contributors for this article: MOM Forward Deployed Team (Barry Tng, Ethan Mak, Joel Koo), and Container Stack team…