ML Engineer & Full Stack Developer From precision mechanics at BMW to high-performance Machine Learning. Focused on optimization, raw performance, and architectural pragmatism.
| Domain | Technologies |
|---|---|
| Languages | Python, Rust, TypeScript, SQL |
| AI / ML | YOLOv8, XGBoost, TATR (Fine-tuning), NLP (CamemBERT), Vector Sim. |
| Backend | Node.js, FastAPI, Ruby on Rails, PostgreSQL, Redis |
| Infrastructure | vst-cli (Pragmatic Deployment), Docker, GitHub Actions, Nginx, Linux |
| Curently studying JEPA Architecture. |
The Pragmatic Deployer. A CLI tool built to simplify VPS management and deployments without the cloud-native overengineering.
- GitHub:
vst-cli- Focus: Automation, low footprint, total infrastructure control.
High-Density Document Extraction. Multi-stage AI pipeline for complex table extraction ("Page-to-Data").
- Performance: RAM Peak
$4.5\text{GB} \rightarrow < 1\text{GB}$ (Rust/Tauri migration).- Logic: Human-in-the-loop verification UI with confidence-based scoring.
Equine Biomechanics SaaS. Real-time postural analysis via Computer Vision and LLM integration.
- Core: YOLO Pose Estimation + Vector Algebra + LLM Feedback Engine.
- Impact: Automated linguistic coaching based on articular angle extraction.
3D Print ERP. Industrial fleet management based on real-time G-Code metadata parsing.
- Business Logic: Live machine depreciation tracking (e.g., 5000h lifespan) and predictive stock management.
[2024-PRESENT] > Independent ML & Software Engineer
> Lead Architect on InsureflowAi & vst-cli.
> Specialized in complex data extraction and technical R&D.
[2023-2024] > Full Stack Dev & Data Analyst @ Hookto (Xefi Group)
> ERP Architecture, Micro-services, Business Intelligence.
[2018-2021] > Expert After-Sales Technician @ BMW Bavaria Motor
> High-level system troubleshooting & precision diagnostics.- Location: (42), France
- Email: pauwelstanguy@protonmail.com
- Portfolio: pauwelstanguy.fr