class Fazilkhan:
role = ["Founder", "Product Architect", "AI/ML Engineer"]
ventures = {
"🏥 NuroVed": "Healthcare Infrastructure → v1 Active",
"🎓 Educle": "EdTech + AI Credibility → Building",
"🌱 Trashee": "CleanTech Community → Building",
"🧠 Actora": "Behavioral Design AI → Building"
}
philosophy = "Build systems that change behavior, not just screens"
metrics = {"patterns_tested": "10,000+", "ops_uplift": "78%"}
def current_focus(self):
return "NuroVed v1 → Real-world usability, not demo polish"| ✅ Optimize For | |
|---|---|
| Output | Outcome |
| Downloads | Daily Active Usage |
| Demo Polish | Real-world Resilience |
| Vanity Metrics | Behavior Change |
┌─────────────────────────────────────────────────────────────┐
│ 01 → Why do users drop off here? │
│ 02 → How does this scale at 10x? │
│ 03 → Where does AI add real signal — not noise? │
└─────────────────────────────────────────────────────────────┘
| Project | Tech Stack | Impact |
|---|---|---|
| 🤚 Air Drawing & 3D Canvas | MediaPipe, OpenCV, Depth Estimation | Gesture-to-drawing without hardware |
| 🩺 Health Record Analyser | NLP, PyTorch, PDF Parsing | Multi-visit clinical pattern detection |
| 🔬 AI Symptom Analyser | Medical Ontologies, Model Inference | 10,000+ disease patterns validated |
| 🥗 AI Diet Planner | PyTorch, Personalization Engine | 500+ dietary scenarios handled |
| 🧪 Food Product Scanner | NLP, OCR, Ingredient Graph | Real-time offline label analysis |
| 🤖 MAVIS AI Assistant | LLM, Context Management | Multi-intent automation engine |
| 📊 Pharmacy Dashboard | React, FastAPI, Supabase | 78% operational uplift |
🔍 View Detailed System Architecture
🤚 Air Drawing & 3D Interactive Canvas
Type: Computer Vision System
Tech: MediaPipe, OpenCV, Depth Estimation, Custom Gesture Classifiers
Innovation: Extended 2D canvas into full 3D interactive space
Status: Production Ready🩺 Health Record Analyser
Type: NLP Clinical Engine
Tech: PyTorch, PDF Parsing, Clinical Pattern Detection
Features: Multi-visit context awareness, Anomaly flagging
Status: Deployed🔬 AI Symptom Analyser
Type: Diagnostic AI Layer
Validation: 10,000+ disease patterns
Approach: Symptom co-occurrence modeling beyond keyword matching
Status: Active Inference| Period | Role | Company | Impact |
|---|---|---|---|
Present |
Product Architect | Promacle | Full lifecycle ownership System design → Production |
2023 |
Lead Developer | Dprofiz Ltd | IoT Waste Management Reward systems architecture |
2022 |
AI/ML Engineer | MiroFish AI | End-to-end AI pipelines Context-aware automation |
@@ NuroVed v1 → Real-world usability as the benchmark @@
@@ Educle → AI-driven student credibility layer @@
@@ Actora → Behavioral design that makes decisions stick @@
@@ AI behavior loops → How model outputs shape product retention @@
@@ System design → Distributed architecture & failure modes @@









