Building SwasthaX: Engineering a Production-Scale AI Platform for Healthcare
summary
Production AI platform that builds a behavioral digital twin from wearables, medical records and daily habits, calculates a prevention score and issues live nudges (e.g., change 10k‑step goal to 6k, alter meal plan) and can auto‑book NHS appointments. Architecture uses LangChain + LangGraph orchestration over Gemini and IBM WatsonX models (moving toward Claude), PostgreSQL for persistence and blockchain for data consent. Early deployment runs in the UK with chikitsa.io in three countries, reported model accuracy ~82% with a >95% target.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 1:27Reactive healthcare limits
- 5:18SwasthyaX architecture
- 12:02Daily health routines
- 16:20Behavioral digital twin
- 21:11Technical stack
- 23:36Health data privacy
- 27:27Deployment status
- 30:39AI healthcare principles
related talks
07mHow AI is Enabling Healthcare's Biggest Blindspot
23mEngineering AI-Powered Predictive Healthcare Systems
25mAI for healthcare: Scaling Access and Quality of Care for Everyone
20mFrom promise to impact with gen AI in healthcare & life sciences
47mLeveraging AI for Global Health
59mArtificial intelligence in Healthcare
27mGen-AI at Scale: Simplifying Orchestration of Healthcare Applications Across Multi-Cluster Environme
46mStream-enabled chronic disease management at population scale