MEMS

Personalized Ambient Intelligence for Alzheimer's & Dementia Care


Introduction

MEMS is an AI-powered personalized ambient-intelligence system that learns the longitudinal behavioral trends, affective inclinations, interaction patterns, and overall personality profiles of elderly members affected by Alzheimer's Disease. It then integrates these insights with behavioral objectives defined by family members and caregivers to target key challenges in Alzheimer's and dementia care management—such as mealtime distress, sundowning, and overnight insomnia. Addressing these issues can significantly improve member quality of life and ease the overwhelming emotional burden caregivers face around the clock.

How It Works

1

Continuous Learning

  • Monitors the member's day-to-day behavioral tendencies, activity preferences, and stress triggers over time.
  • Interprets and predicts behavioral trajectories across more than 50 unique measures of core affect, enabling discovery of subtle deviations in mood and real-time forecasting of significant emotional episodes.
2

Caregiver & Family Objectives

  • Behavioral objectives consist of key tasks assigned to MEMS that address the most commonly encountered pain points in Alzheimer's care management.
  • Examples of objectives offered include standardizing meal timing, mitigating agitation in sundowning members, and fostering high-quality sleep to re-enforce member circadian rhythms.
  • MEMS works to answer these objectives in real time and adapt its recommendations accordingly.

Interface

Please contact us directly for details on our AI model’s architecture and operation. Detailed technical specifications are not shared.

Example in Action

For key moments, caregivers or family select an objective—such as optimal mealtime—from a checklist. Once selected, MEMS autonomously manages the process, continuously tracking behavioral signals, analyzing context, and delivering precise, real-time prompts through intuitive light patterns.

Scenario Setup

It's mid-morning at home, and lunch is coming up. The user selects the mealtime – when to eat objective from the checklist.

MEMS Analysis

MEMS assesses the member's current affective state, compares it to their context-aware baseline, and determines the optimal time to recommend a meal.

Recommendation

A green light appears, indicating an optimal window to suggest the member that it is time to eat.

Result

Mealtime proceeds smoothly, with fewer refusals and more positive engagement.

Validated Impact

Main Validation (2025)

Main validation of both the hardware product and initial models has seen the greatest success at home in Florida, supporting:

  • Families caring for elderly members with Alzheimer's or dementia
  • Early-stage members seeking more independence

Additional Pilots (2024)

Additional pilots have been run in three centers across California, Florida, and Asia to:

  • Expand data collection for training model weights and improving adaptability across diverse elderly populations
  • Access a greater variety of members and care settings to strengthen real-world performance
  • Operate at breakeven costs to make participation easy for facilities and encourage collaboration
  • Lay the groundwork for the future at-home device, where we see the largest long-term impact and scalability

These pilots have generated measurable improvements in emotional stability, routine smoothness, and caregiver stress reduction. Please contact us for validation methodology and quantitative pilot results.

Why the At-Home Version Has Highest Impact

For families and in-home caregivers, MEMS directly eases the daily emotional and decision-making load:

Reduces Guesswork

Turns uncertainty into clear, timely prompts.

Eases Emotional Burden

Anticipates issues before they escalate.

Supports Meaningful Moments

Highlights the best times for connection so families can focus on joy, not just logistics.

Gives Caregivers Breathing Room

Handles small but critical moments that make daily life smoother.

Core Team

Victor Yin

Co-Founder – victoryin@memsdev.com

M.S. Computer Science (AI), B.S. Computer Science (AI), B.A. History – Stanford University

Mark Laurie

Co-Founder – marklaurie@memsdev.com

Ph.D. Candidate, Computational & Mathematical Engineering; M.S. Computer Science (AI); B.S. Biomedical Computation – Stanford University

Additional Team – Biostatistics PhD based at Stanford for model validation, Audio sentiment analysis built by PhD based at UC Berkeley, Hardware designer based in Sweden.

Ask

We recently raised a small friends & family round and are preparing for our next stage of growth. We're eager to meet more people through introductions that can help us sharpen our direction and expand our reach.

We are seeking:

New pilot partners in both home and facility settings

Strategic collaborators to facilitate marketing and distribution.

Investment and grant funding to accelerate deployment

Our ultimate vision is to make MEMS a household standard for Alzheimer's and dementia care, seamlessly connecting home and facility environments into a continuous care network.