👋 Hi, I'm J.P Mukhwana

Edge ML Engineer | Low-Resource ML Optimizations | Building AI for the 1.3B people computing on 2G

┌─────────────────────────────────────────────────────┐
│ 🌍 I solve problems others choose to forget         |
│                                                     │
│ While others optimize for GPU clusters,             │
│ I optimize for reality:                             │
│ • $50 Android phones from 2015                      │
│ • 2G internet, 512MB RAM                            │
│ • Pay-per-kilobyte data                             │
│ • 1.3 billion people at the edge                    │
└─────────────────────────────────────────────────────┘

🚀 Current Project: Epital Elderguard

Production-ready ML models optimized for African infrastructure

[Explore Epital Elderguard] →](https://github.com/petitmj/-Epital-ElderGuard---AI-Powered-Fall-Detection)

---

📊 Impact in Production

╔══════════════════════════════════════════════════════╗
║  40+ ML models deployed on Edge Devices              ║
║  2.4M+ end users (95% on <2GB RAM devices)           ║
║  $4K+ saved in cloud costs                           ║
║  <100ms average inference latency                    ║
║  80% of models work completely offline               ║
╚══════════════════════════════════════════════════════╝

---

🔧 Tech Stack

Edge ML & Optimization:
TensorFlow Lite ONNX Runtime PyTorch Mobile Quantization Pruning Distillation Federated Learning ARM Optimization

Languages:
Python JavaScript/TypeScript C++ Rust PHP

Frameworks:
TensorFlow PyTorch scikit-learn Django React Next.js WordPress

Infrastructure:
Linux/Ubuntu Docker Kubernetes Edge Computing IoT CI/CD

---

📝 Latest Articles

View all articles →

---

🎯 Featured Projects

Epital Elderguard Real-time fall-detection system for Elderly Patients 92% fall-detection accuracy (hybrid model) | 4.3ms average on-device inference | <10s average alert delivery Code • Demo • Article

---

📊 GitHub Stats:

---

---

💼 Currently Seeking

Remote roles where I can:

• ✅ Build ML systems for global users (not just SF)
• ✅ Optimize for constraints (better engineering)
• ✅ Write docs people actually understand
• ✅ Contribute to meaningful open source

Target: Companies serious about "democratize [X]" globally
Automattic • Zapier • Canonical

---

📫 Let's Connect

---

"The constraints of building for Africa make me a better engineer for everyone. I have learnt to respect bandwidth, memory, and compute. I ship lean, efficient, elegant solutions."

```