Field Service Engineer turned AI/ML Developer. I build things that solve real problems—from predictive models for polymer design competitions to full-stack fleet management systems. My background in medical imaging hardware gives me a unique perspective on system reliability and data integrity.
🧪 NeurIPSOpenPolymerPrediction2025 ⭐2
Machine learning pipeline for polymer property prediction. Competition submission for NeurIPS 2025.
Python Scikit-learn Feature Engineering
🤖 RAG_Showcase
Retrieval-Augmented Generation implementation demonstrating context-aware LLM responses.
Python LangChain Vector Databases
📊 IRA_showcase ⭐1
Insurance Risk Prediction Application
Python Data Processing
🚗 fleet_mgmt_django ⭐1
Full-featured vehicle fleet management system with real-time tracking and maintenance scheduling.
Django PostgreSQL REST APIs
🔧 universal_solver
Modular problem-solving framework using computational notebooks.
Jupyter NumPy Scientific Computing
Languages: Python • JavaScript/TypeScript • SQL
AI/ML: PyTorch • Scikit-learn • LangChain • OpenAI API • Hugging Face
Web: Django • React • Node.js • PostgreSQL • REST APIs
Tools: Git • Docker • Jupyter • Linux • CI/CD
Spent years keeping CT scanners, C-arms, and mobile X-ray systems running in high-stakes environments. That taught me two things: complex systems fail in interesting ways, and good diagnostics beat guesswork every time. Now I apply that same debugging mentality to ML models and production code.
Competed in data science challenges (NeurIPS, Ariel Data Challenge) because I learn best by shipping real solutions under constraints.
- Advanced transformer architectures and their practical applications
- Production ML systems and model deployment patterns
- Building robust data pipelines for real-world datasets
📧 jedrzej.grabala@gmail.com
🌐 jgitsolutions.space
☕ Buy me a coffee
Open to collaboration on ML projects, technical discussions, or just talking shop about system design and data engineering.