AI Engineer building real-world applications with LLMs, computer vision, and full-stack systems
I design and deploy end-to-end AI products — from model development to production-ready apps.
Currently building FormCheck AI, a real-time exercise form detection app using computer vision and deep learning.
📍 Los Angeles, CA · Open to AI / Applied AI roles
- Build AI-powered applications (LLMs, RAG, multi-agent systems)
- Develop real-time computer vision systems
- Ship full-stack products (React + FastAPI + APIs)
- Focus on practical, production-ready AI — not just notebooks
Real-time exercise form detection using MediaPipe + LSTM
- Detects 30+ faults across barbell exercises
- Real-time feedback + post-set analysis
- Built with Python, TensorFlow, React Native, FastAPI
RAG-based chatbot with intelligent query routing
- Multi-agent architecture (search / genre / analysis)
- FAISS vector search over 2,700+ films
- 98% plot coverage via multi-API enrichment
Machine learning system on real financial data
- XGBoost model with ROC-AUC: 0.90
- Improved recall via threshold tuning
- Handled class imbalance with SMOTE
AI/ML: Python, TensorFlow, scikit-learn, LLMs, RAG, LangChain
Backend: FastAPI, Node.js, APIs
Frontend: React, React Native
Data: FAISS, Pandas, NumPy
Tools: Git, Vercel, Supabase
- GitHub: https://github.com/kamilj62
- Email: kamilj@umich.edu
Build things that work in the real world — not just in notebooks.