Doctor Saheb
Full-stack healthcare platform featuring LLM-assisted symptom guidance, Spring Security JWT auth, MongoDB, and Caffeine caching.
Backend Architect & Security Lead
2 Developers
Healthcare Engineering Project
Healthcare & Backend Architecture
December 2025
JWT + Caffeine Caching
The Problem & Real-World Friction
Patients needing quick health guidance often face delayed triage and lack secure, authenticated platforms to track consultations.
Architectural & Technical Approach
Built a full-stack platform featuring LLM-assisted symptom guidance with Groq AI, secured via Spring Security and JWT authentication, backed by MongoDB and Caffeine in-memory caching for low latency.
Key Technical Highlights
AI Healthcare Guidance & Spring Security System
Sub-Second Groq AI Inference
High-speed symptom guidance engine powered by Groq Llama inference providing fast initial medical recommendations and triage.
Stateless Spring Security & JWT
Role-based access control protecting patient consultation history and doctor appointments with cryptographically signed JWT tokens.
Multi-Layer Caffeine Caching
In-memory Caffeine cache layer caching high-frequency symptom databases to ensure sub-10ms response times for common queries.
MongoDB Document Persistence
Flexible NoSQL schema accommodating dynamic consultation notes, prescription structures, and polymorphic medical history logs.
Project Results & Impact
Delivered secure, cached medical symptom triage with sub-second API response times.
Reproducible Benchmarks & Test Setup
Test setup, evaluation procedure, and validated outcomes reported for this architecture:
December 2025
Apache Bench (ab -n 1000 -c 50) testing cached specialist directory endpoint.
Baseline: Direct un-cached database query: 68ms avg.
Outcome: Caffeine cache hit latency: 6.2ms avg; 94.8% cache hit ratio under concurrency.
Separation of Implemented, Prototyped & Planned Features
- •Stateless Spring Security JWT authentication and role-based access
- •Caffeine in-memory cache layer for specialist directories and common symptoms
- •MongoDB schema for encrypted consultation notes
- •Groq Llama 3 70B symptom guidance prompt pipeline
- •Telemedicine WebRTC encrypted video consultation room
Role & Contributions
Backend Architect & Security Lead
2 Developers
- Designed stateless Spring Security filter chain with HMAC-SHA256 JWT validation.
- Implemented Caffeine in-memory cache layer with W-TinyLFU eviction policy for specialist directories.
- Created MongoDB medical consultation persistence layer and Groq Llama 3 70B symptom triage prompts.
Technologies Used
8 toolsCivic AI
AI-powered civic issue reporting and government management platform that won Champion at the AI & API Hackathon 2026.
View next case studyInterested in this architecture?
Have a question about the system design, tech stack, or potential collaboration?
Let's discuss