HACKaSTONE 2027

Challenges

Published 6 October 2026

CH-001Provided by A*STAR

H&E-based Digital Immune Reporter

Develop AI models that infer spatial immune profiles within the tumor microenvironment from routine H&E histopathology images. The aim is to improve patient stratification and predict responses to immunotherapy through computational characterization of the tumor immune microenvironment.

CH-002Provided by A*STAR

Agentic AI-Enabled Bioinformatics Platform for High-Plex Tissue Imaging

Develop an agentic AI-enabled bioinformatics platform that automates and integrates the analysis of high-plex tissue imaging data, enabling cell phenotyping, spatial profiling and biologically interpretable discoveries within the tissue microenvironment.

CH-003Provided by NTU

Hotline Agent for Domain-Specific Telephone Service

With labour costs rising, most organisations have replaced manual hotline service with chatbots. These run 24/7, but they are not intelligent and lack a human touch. This challenge is to design an AI agent for hotline and front-desk telephone service. Domain-specific knowledge will be used to train the agent, with the potential for self-improvement.