Faiaz Amin Khan

I research and build systems that span the physical-digital divide. From designing patent-grade acoustic sensing mechanism at Samsung R&D to deploying AI workflows in resource contrained device, my work focuses on decoding human signals to create seamless, intelligent interactions.

My research focuses on the full stack of human augmentation: designing novel sensing mechanisms that capture physiological signals and building agentic machine intelligence workflows that decode those signals into proactive health and interaction insights.

Core Research Themes

1. Multimodal Wearable Sensing

Fusing IMU, PPG, acoustic, and contextual data to model human activity, behavior, and physiology in the wild.

  • Activity-based health insights on Galaxy Watch (IMU + PPG + context)
  • Sensor data fusion for robust, noise-tolerant inference

2. Acoustic and Novel Input for Wearables

Designing acoustic-sensing input mechanisms and subtle interaction channels for resource-constrained devices.

  • Low-intensity acoustic input for headless wearables (MFCC + spectrogram)
  • Homonym disambiguation based on tone and prosody for hands-free control

3. AI for Mobile Health & Personal Informatics

Using time-series modeling and context-aware feedback to support health behavior change and self-reflection.

  • Turning sensor streams into interpretable health insights
  • Personalization across users, routines, and environments

4. On-Device & Resource-Constrained ML

Optimizing models, pipelines, and sensors to run directly on watches, phones, and IoT devices.

  • TensorFlow Lite deployment
  • Model compression and pipeline optimization for real-time operation

5. Learning & Optimization

Reinforcement learning and curriculum learning for structured problems and adaptive systems.

  • RL-based TSP with hardness-adaptive curriculum
  • Ideas toward adaptive intervention policies for health systems