class Fazilkhan:
role = ["Founder", "Product Architect", "AI/ML Engineer"]
ventures = {
"π₯ NuroVed": "Healthcare Infrastructure β v1 Active",
"π Educle": "EdTech + AI Credibility β Building",
"π± Trashee": "CleanTech Community β Building",
"π§ Actora": "Behavioral Design AI β 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 β Real-world usability, not demo polish"| β Optimize For | |
|---|---|
| Output | Outcome |
| Downloads | Daily Active Usage |
| Demo Polish | Real-world Resilience |
| Vanity Metrics | Behavior Change |
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
β 01 β Why do users drop off here? β
β 02 β How does this scale at 10x? β
β 03 β Where does AI add real signal β not noise? β
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
| Project | Tech Stack | Impact |
|---|---|---|
| π€ Air Drawing & 3D Canvas | MediaPipe, OpenCV, Depth Estimation | Gesture-to-drawing without hardware |
| π©Ί Health Record Analyser | NLP, PyTorch, PDF Parsing | Multi-visit clinical pattern detection |
| π¬ AI Symptom Analyser | Medical Ontologies, Model Inference | 10,000+ disease patterns validated |
| π₯ AI Diet Planner | PyTorch, Personalization Engine | 500+ dietary scenarios handled |
| π§ͺ Food Product Scanner | NLP, OCR, Ingredient Graph | Real-time offline label analysis |
| π€ MAVIS AI Assistant | LLM, Context Management | Multi-intent automation engine |
| π Pharmacy Dashboard | React, FastAPI, Supabase | 78% operational uplift |
π View Detailed System Architecture
π€ 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 Readyπ©Ί Health Record Analyser
Type: NLP Clinical Engine
Tech: PyTorch, PDF Parsing, Clinical Pattern Detection
Features: Multi-visit context awareness, Anomaly flagging
Status: Deployed㪠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 β 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 β Real-world usability as the benchmark @@
@@ Educle β AI-driven student credibility layer @@
@@ Actora β Behavioral design that makes decisions stick @@
@@ AI behavior loops β How model outputs shape product retention @@
@@ System design β Distributed architecture & failure modes @@









