VitalSync — Digital-Twin Elderly Care
ESP32 wearable sensors + 91.8% risk prediction, 95.1% fall detection, SHAP explainability, and symbolic safety guardrails.
The situation & problem
Elderly home monitoring currently suffers from disconnected panic buttons, uncalibrated sensor drift, and alarm fatigue. Clinicians and family caregivers lack a continuous predictive digital twin that contextualizes vitals with historical baselines.
The operational constraints
Resource-constrained edge hardware (ESP32 MCU, MAX30100 pulse oximeter, MPU6050 accelerometer/gyroscope, Neo-6M GPS). Telemetry ingestion contracts required 99.4%+ integrity over lossy wireless connections, sub-second emergency dispatch, and verifiable, grounded LLM health memory.
Architectural tradeoffs & what was refused
A generic wearable dashboard streaming unvalidated JSON. A consumer voice assistant hallucinating ungrounded medical advice. Uninterpretable black-box alert thresholds.
What I built & shipped
End-to-end digital twin system: ESP32 edge firmware with Kalman filtering; FastAPI ingestion pipeline with strict schema validation; Next.js real-time caregiver portal; 91.8% health-risk prediction engine with SHAP feature attribution; 95.1% accuracy kinematic fall-detection model; conversational memory bot backed by deterministic symbolic safety guardrails.
Architecture & stack
ESP32, C++, Next.js, TypeScript, Tailwind CSS, FastAPI, Python, PostgreSQL, Supabase, Scikit-learn, SHAP, LangChain.
Result & measurable impact
Achieved 91.8% health-risk prediction accuracy, 95.1% fall detection precision, 99.4% telemetry data-ingestion contract integrity, and no ungrounded medical advice observed across [TODO: Insert simulated conversation count N] simulated clinical test conversations.
Next iterations & roadmap
Clinical trial deployment across regional elder-care residency facilities with real-time ambulatory telemetry.
System surfaces & interface captures
Hover and click any record to open the full-resolution pixel inspector with zoom lens and 1:1 precision.
Real-time caregiver monitoring dashboard displaying live vital telemetry, pulse rates, and predictive risk indicators.
Longitudinal telemetry analytics showing 24-hour SpO2 trends, body temperature curves, and Kalman-filtered motion data.
Custom ESP32 micro-controller wearable board integrated with MAX30100 pulse oximeter, MPU6050 kinematic IMU, and GPS module.
End-to-end telemetry and inference pipeline: edge sensor ingestion, FastAPI validation layer, Supabase PostgreSQL store, and Next.js frontend.
SHAP feature attribution waterfall chart explaining the specific biometric features driving the 91.8% health-risk classification.
Real-time caregiver monitoring dashboard displaying live vital telemetry, pulse rates, and predictive risk indicators.
Longitudinal telemetry analytics showing 24-hour SpO2 trends, body temperature curves, and Kalman-filtered motion data.
Custom ESP32 micro-controller wearable board integrated with MAX30100 pulse oximeter, MPU6050 kinematic IMU, and GPS module.
End-to-end telemetry and inference pipeline: edge sensor ingestion, FastAPI validation layer, Supabase PostgreSQL store, and Next.js frontend.
SHAP feature attribution waterfall chart explaining the specific biometric features driving the 91.8% health-risk classification.
Engineered for real operational outcomes.
What founders, engineering leads, and operators say about working directly with me.
“May Allah reward you with good. I was truly fortunate to know you and work with someone so respectful, honest, and conscientious in his craft.”
Suliman H.
Founder & Platform Lead · Souq Al Manakh
Kuwait
“Masha'Allah, may Allah grant you health and strength. The platform looks truly massive, polished, and impressive.”
Eng. Abdulrahman W.
Founder & Lead · Invie (Event Screens & Digital Invites)
Kuwait
“Definitely not our last collaboration. You are a creative engineer who does honest work with high integrity. May Allah bless your career and efforts.”
Founder & Director
Executive Director · Jisr Al-Amal Platform
Saudi Arabia
“Honestly, I was hesitant about hiring online, worried the work would be mediocre. But working with Abdelrahman completely changed my perspective—he delivered outstanding work, showed immense flexibility in handling revisions, and proposed technical solutions that saved me real money. Highly recommended.”
Muhanad A.
Business Client
Saudi Arabia / GCC
“Honored to work with such a trustworthy and professional engineer. Your work was exceptional—the design truly blew me away, and your technical advice for developing and scaling the platform was invaluable. Looking forward to our ongoing projects together.”
Ali S.
Platform Founder
Egypt / Regional
“My sincere thanks for your high professionalism and precision. What truly sets you apart is your continuous follow-through until the very last detail, reflecting genuine dedication to your craft, Engineer Abdelrahman.”
Alaa A.
Business Client
Saudi Arabia
“Extremely talented and genuinely invested in delivering the best outcome with valuable architectural advice. Commendable work, wishing him the best.”
Saad A.
Product Owner
Saudi Arabia
“Thank you Eng. Abdelrahman, you went above and beyond. Excellent technical troubleshooting and code restructuring.”
Mohammed R.
Technical Client
Arab Region
“Thank you for the excellent work. We look forward to our continuous long-term collaboration.”
Omar J.
Business Owner
Arab Region
“Thank you for the rapid turnaround, creative delivery, and flawless execution. Fast communication and completely trustworthy.”
Abdulrahman A.
Marketing Client
Arab Region
“Outstanding service and a highly respectful professional. Thank you very much.”
Daham A.
Product Client
GCC Region
“May Allah reward you, outstanding execution and dedicated delivery.”
Esaam A.
Commercial Client
Arab Region
Have a similar constraint?
Tell me where your current operational system or product is breaking.