HÀ NỘI — The AWS Regional Large Language Model (LLM) League wrapped up its 2025 finals in Singapore, bringing together top university students from six ASEAN countries to demonstrate their AI capabilities and fine-tuning skills.

Co-organised by Amazon Web Services (AWS) and AI Singapore (AISG), the first-of-its-kind regional competition attracted over 1,300 students from Indonesia, Malaysia, the Philippines, Singapore, Thailand and Việt Nam. 

Participants competed by fine-tuning LLMs using Amazon SageMaker JumpStart tools, aiming to outperform a larger 70B model in a quiz-based challenge.

After several national rounds, finalists gathered at the Grand Finale on 29 May, held during the AI Student Developer Conference. Their models were evaluated by a panel of experts, real-time voting and a benchmark powered by Claude 3.5 Sonnet from Anthropic.

Blix Foryasen from the Philippines claimed first prize, using model distillation to beat the larger model in accuracy. Việt Nam’s Kim Seokyung, a software engineering student from RMIT Vietnam, placed second despite being new to LLM fine-tuning. “I went from zero to a fine-tuned model ready for real use in just weeks,” she said.

“The LLM League embodies our commitment to digital inclusion,” said AWS Country Manager at Public Sector Singapore Elsie Tan,. “It shows how accessible AI skills can be when the right tools and mentorship are in place.”

Through hands-on workshops and real-world simulation, the League equipped students with experience in dataset curation, prompt engineering and responsible AI, preparing them for future careers in AI. The success of this inaugural regional edition suggests a scalable model for AI education across diverse education systems and underrepresented communities in ASEAN.

As Generative AI continues to drive digital transformation, initiatives like the LLM League are helping to shape a diverse, inclusive talent pipeline for the region’s future. — VNS