NUS Students Build AI Marking Tool That Slashes Teacher Grading Time by 99%

Three National University of Singapore students have developed an AI-powered marking platform called Ren that dramatically reduces teacher grading time while preserving the quality and personalisation of feedback. The tool, built by Wong Eu En (Year 2 Computer Science), Justin Cheah (Year 4 Computer Science and Business), and Natasha Koh (Information Systems graduate), slashed one senior lecturer's marking time for 490 students from six weeks to just five minutes for the first pass. By the second or third assignment, teachers typically accept 80 to 90 per cent of AI-generated comments without modification, the team reports.

Ren operates on a human-in-the-loop model the founders call the "assistant approach" rather than the "centurion approach" — teachers upload rubrics, syllabus materials, and learning outcomes, the AI generates a draft grade and detailed feedback, but educators must review, edit, and approve every comment before it reaches students. Over time, the system adapts to each teacher's marking style, preferred phrasing, and even punctuation conventions. Students can submit handwritten or typed assignments, and the platform generates class-wide analytics highlighting strengths and weaknesses by topic and question type after each marking round.

The timing is significant. A recent survey found that three in four Singapore teachers now use AI — more than double the OECD average — and the Ministry of Education has been rolling out its own tools including Markly for feedback generation and an Authoring Copilot for lesson planning on the Student Learning Space platform. Ren's founders are positioning their tool as a complement to these official efforts, focusing on the specific pain point of marking consistency across different teaching assistants and the overwhelming volume of student work in large classes. Senior Lecturer Lee Boon Kee noted that the tool "can systematically check whether students applied several ethical frameworks and give detailed feedback," solving the long-standing problem of inconsistent grading across different markers.

Ren is currently piloting at 11 institutions including NUS and the School of Science and Technology, projected to serve about 5,600 students and work with 50 educators by July. The startup has partnered with self-help group Mendaki to provide free A-level tutoring support for about 600 students in literature, history, chemistry, and economics. The first full schoolwide rollout is scheduled at St Andrew's Junior College in July, with schools subscribing via customised annual plans. Co-founder Wong Eu En explained that the team's philosophy is straightforward: "AI assists, teachers decide."

Why it matters for Singapore: Singapore's education system has been under pressure as teacher workload complaints and burnout rates made headlines through 2025 and 2026. Tools like Ren represent a concrete answer to the question of how AI can genuinely reduce administrative burden rather than adding another layer of tech to manage. If the pilot results hold, AI-assisted marking could become a standard fixture in Singapore classrooms, freeing teachers to focus on the students who need the most help while maintaining the high feedback standards that Singapore's education reputation depends on. The fact that three undergraduates built this — and that 11 institutions are already piloting it — says as much about Singapore's talent pipeline as it does about the product itself.