Discover our initiatives in Optimised Computing and High-Performance Systems
Projects focused on improving and optimising enterprise-level computer architecture, covering both hardware and software.
Leveraging high-performance computing to accelerate molecular docking simulations, enabling efficient screening of herbal formula compounds for potential anti-asthma therapeutic candidates.
Twin-B is a digital twin platform that simulates occupant behavior inside buildings by coupling EnergyPlus building energy simulation with Mesa agent-based modeling, powered by High-Performance Computing on the LANTA supercomputer.
A proof-of-concept NLP chatbot designed to preserve local culinary knowledge while learning from interactions with international visitors, addressing Chiang Mai's 105%+ annual hospitality staff turnover through persistent knowledge repositories and culturally-aware service recommendations.
A proof-of-concept co-simulation system combining agent-based modeling and large language models to generate intelligent, culturally enriched EV tourism routes — addressing Northern Thailand's 60–100km charging gaps and transforming range anxiety into confident exploration.
An HPC-powered platform that parallelizes atmospheric PM2.5 transport models, real-time environmental sensor networks, and AI-driven wellness personalization — achieving 92.5% forecasting accuracy and converting Lampang's air quality challenge into a competitive wellness tourism advantage.
Born from student volunteer experiences during Thailand's worst flooding in 80 years, ICS NEXUS integrates HPC flood simulation on LANTA, AI/LLM multi-dialect warning translation, and real-time GIS — targeting 80%+ prediction accuracy and comprehensible warnings for 91,894 at-risk households.
Deploying the LANTA supercomputer to enable real-time Thai-Hmong translation — reducing processing from 15–20 seconds to 3.7 seconds with 82.3% accuracy, bridging healthcare, education, and agricultural communication barriers for 200,000 Hmong speakers across Northern Thailand.
Utilizing LANTA's 704 A100 GPUs to develop culturally-aware safety assessment for San Kamphaeng's unique architecture — achieving 100x speedup for real-time hazard detection that respects Buddhist altars and spirit houses while identifying genuine risks in workshop-home environments.
An HPC-powered livelihood matching system for Thailand's poorest province — overcoming linguistic barriers across 6+ ethnic languages to connect elderly residents with income-generating opportunities, preserving indigenous knowledge while targeting poverty reduction in skipped-generation households.
Deploying HPC to train and compress multi-language NLP models by 90%+ for Raspberry Pi edge deployment — enabling offline disaster warnings in Thai, Northern Thai, and hill tribe languages with <2-second translation, field-tested during the 2024 Northern Thailand floods.
An HPC-powered computer vision system democratizing expert longan pruning knowledge via smartphone photography — targeting premium-grade production increase from 25% to 65%, reducing disease losses from 35% to 10%, and building Thailand's first 1M+ annotated longan image repository.