The One About AI x Healthcare (Round VI)
What does it take to move AI from promising healthcare research into clinical practice?
From personalising drug dosing for individual patients to detecting cancer from digital pathology slides, AI is being developed for very different parts of the care journey. Join researchers and practitioners as they share the technical approaches behind these systems, how they're developed and validated, and the challenges of bringing AI into clinical workflows.
More About the Sharings
Peter Wang (Research Fellow, NUS) will present “CURATE.AI: Personalising Drug Dosing for Individual Patients”
CURATE.AI is an AI-enabled, small data-driven platform that dynamically optimises drug dosing using only patient-specific data. This presentation will highlight several CURATE.AI trials and discuss key considerations for the future implementation of AI-enabled technologies in clinical workflows. (Technical Difficulty: 200)
Yusuke Minami (Senior AI Researcher, Qritive) will share on "Building AI for Lymph Node Metastasis Detection"
Pathologists routinely analyse complex tissue samples to diagnose cancer, but translating this expertise into reliable AI systems presents a different set of challenges. Yusuke will share how Qritive develops deep learning models to assist in detecting lymph node metastasis from digital pathology slides.
Follow the development pipeline from processing large whole-slide images to metastasis segmentation and classification, and learn how these models are validated for use in digital pathology. Through Qritive's work on lymph node metastasis, Yusuke will also discuss some of the practical considerations behind developing computer vision systems for clinical applications. (Technical Level: 100)
Dr Ankur Sharma will share on "GenomicsCopilot: Agentic AI for Variant Interpretation"
Every patient's genome comes back with dozens of variants nobody's sure about yet. Right now a human has to manually check ClinVar, population databases, and prediction scores for each one, which doesn't scale. I built GenomicsCopilot to automate that process: you give it a VCF file and the patient's symptoms, and it comes back with ranked variant calls, each one showing exactly what evidence it used and which rule it applied. The code handles the lookups and math, the AI only handles judgment calls, and I added a second AI system that independently reviews every call over MCP. I'll walk through the architecture, run it live against real genetic databases, and show a case where my system and the second reviewer actually disagreed on a variant, and why that disagreement is useful information rather than something to hide. (Technical Level: 200)
More About the Speakers
Peter Wang is currently a Research Fellow at The Institute for Digital Medicine (WisDM), National University of Singapore. His research focuses on developing AI-enabled digital medicine platforms to address global healthcare challenges spanning infectious diseases and healthy aging. He was directly involved in the first human validation of CURATE.AI, and has also contributed to the development of an AI-enabled pandemic preparedness workflow that rapidly designs drug combinations against infectious diseases.
Yusuke Minami is a Senior AI Researcher at Qritive, where he develops AI models for lymph node metastasis detection. He has five years of experience building end-to-end AI solutions, with previous work spanning image-based fraud detection, recommendation systems, and vector search using product images and user behaviour.
Ankur Sharma has a PhD and 8+ years building production ML systems, based in Singapore. Recent work includes fine-tuning genomic foundation models, building explainable ML classifiers, and multi-agent AI systems, all open source on GitHub.
More About The Series
AI Wednesdays is Lorong AI’s weekly gathering, bringing together practitioners, researchers and innovators for technical discussions on research insights, product development and engineering practices.
Get involved: Learn more about Lorong AI | Speaker Sign-up | WhatsApp Community | LinkedIn | X