Academic
PhD Candidate · National University of Singapore
Building user-aware AI systems for adaptive, human-centered wearables.
I am a PhD candidate at the National University of Singapore, working with Prof. Suranga Nanayakkara in the Augmented Human Lab.
My research sits at the intersection of Human–Computer Interaction and Computer Vision, where I explore how intelligent systems can understand people through multimodal cues, interpret user context, and adapt their responses in real time. Ultimately, I aim to create wearable systems that feel less like tools and more like supportive partners.
More broadly, I enjoy building data-driven systems, spanning ML modeling, prototyping, and deployment that improve decision-making and everyday user experiences.
Open to collaborations and internship opportunities
Interested in applied AI, HCI, XR, and wearable technology projects where multimodal sensing can support better real-time decisions.
News
- July 2026 – “HeadRoom” was conditionally accepted to UIST 2026.
- Apr 2026 – Presenting our workshop paper at the ABAW workshop at CVPR 2026. See you in Denver!
- Mar 2026 – Won 3rd place in the EMI Challenge at the ABAW workshop at CVPR 2026.
- Feb 2026 – Our paper “Interrupting Autopilot” was accepted to CHI LBW 2026. See you in Barcelona!
- Jan 2026 – Our literature review on user-aware adaptive assistive wearables is in review at ACM Computing Surveys.
- Aug 2025 – Gave a talk on Adaptive Cognitive Augmentations at Meta Reality Labs, Redmond, Seattle.
- Feb 2025 – Our paper “VRSense” was accepted to CHI LBW 2025. See you in Japan!
- Sep 2024 – Started an internship via AHLab × Meta Reality Labs collaboration.
- Sep 2024 – Passed my Qualifying Examination. Now officially a PhD Candidate.
- Aug 2023 – Began my PhD at the National University of Singapore (NUS).
- Jan 2023 – Started Data Analytics consulting at LIRNEasia.
- May 2022 – Joined Axiata Digital Labs as a Data Engineer.
- Aug 2022 – Graduated with First Class Honors (Electronic & Telecommunication Engineering), University of Moratuwa.
- June 2022 – “CrossPoint” presented as a full paper at CVPR 2022.
- Oct 2022 – “3DLatNav” presented as a workshop paper at ECCV 2022.
Affiliations
Industry
Research Projects
User-Aware Adaptive Assistive Wearables
We conducted a systematic literature review of 63 papers examining how adaptive wearables sense user states, trigger context-aware interventions, and support real-time cognitive or behavioral feedback. We introduce a taxonomy of sensing modalities, adaptation triggers, and intervention strategies, and highlight key design challenges and opportunities. This work provides a foundation for developing next-generation wearables that meaningfully adapt to users' needs. This SLR is currently under review at ACM Computing Surveys.
HeadRoom: Lightweight, Edge-deployable Pipeline for Adaptive Notification Routing
HeadRoom is a lightweight, low-latency pipeline that estimates visual and auditory channel availability from egocentric video and audio, then routes notifications to the sensory channel that is more available in the moment. In a controlled user study (N=25), we found that routing notifications to the more available channel reduced response time under high perceptual load compared with routing them to the less available channel. The pipeline is designed for real-time deployment on wearable and mobile edge hardware. Project website.
Sensory Spotlight: Anticipating User Attention
Sensory Spotlight explores how AI can anticipate shifts in human attention by combining audio and visual signals, similar to how we react to salient events in our environment. The model predicts attention shifts and provides saliency to indicate which modality "got the spotlight," supporting decisions about where and how feedback should appear (e.g., smart glasses display vs. audio). Part of this project is ongoing, we have very interesting results!
VRSense: An Explainable System to Help Mitigate Cybersickness in VR Games
VRSense is an explainable system to help VR game developers assess cybersickness. Instead of a black box, the system uses interpretable features to provide actionable insights into game design and user interactions. Designed to be plug-and-play, VRSense helps developers understand how effectively their game mitigates motion sickness. Read the paper.
3D Object Transformation and Regeneration for Privacy in Mixed Reality
We developed a 3D-2D correspondence technique for point clouds and a 3D vision algorithm that lets users add, delete, or modify parts of a 3D object, enabling regeneration on the other end. Evaluated against simulated privacy attacks and implemented on-device for real-world feasibility. Resulted in two papers: CrossPoint (CVPR) and 3DLatNav.
Work Experience
Research Intern - Meta Reality Labs x Augmented Human Lab
Led multimodal XR user studies and predictive modeling for real-time cybersickness assessment.
- Led the design and execution of a large-scale VR gameplay user study (N=150), investigating motion sickness during active VR interactions.
- Collected and synchronized multimodal high-frequency data including eye tracking, head motion, and physiological signals in real-world XR settings.
- Developed machine learning models to predict real-time discomfort and cybersickness onset during gameplay.
- Built robust data pipelines for aligning heterogeneous sensor streams at scale.
- Translated model outputs into actionable insights for VR game evaluation on VR gaming platforms.
Data Engineer - Axiata Digital Labs
Built and productionized ML pipelines across telecom domains on cloud infrastructure.
- Engineered scalable ML/DL deployment pipelines for the company's AI Factory platform, supporting model development, testing, and production integration.
- Designed and maintained data ingestion and model integration workflows across multiple telecom business domains.
- Developed and deployed a customer churn prediction model with end-to-end pipelines, enabling data-driven decision making for enterprise clients.
- Worked across AWS and GCP environments to support reliable, production-grade ML systems.
- Collaborated with product and engineering teams to bridge research prototypes and real-world business applications.
Data Analytics Consultant - LIRNEasia
Delivered geospatial ML workflows and longitudinal datasets to support policy-focused analytics.
- Designed and implemented ML pipelines for classifying built-up regions from satellite imagery to support a national-scale urbanization index.
- Integrated geospatial workflows using QGIS and Google Earth Engine for automated spatial analysis and visualization.
- Built longitudinal datasets to study household energy consumption patterns across regions and time.
- Collaborated with policy researchers to translate data-driven findings into evidence-based insights for urban planning.