Manufacturing Data Scientist — IIoT · Predictive Maintenance · LLM Pipelines · Aerospace
4+ years building ML and LLM systems on industrial sensor data across laser welding, CNC milling, and additive manufacturing at Laseronics LLC (AS9100 / ITAR aerospace laser shop).
| Project | What it does |
|---|---|
| manufacturing-log-analyzer | LLM diagnostics pipeline for machine logs — anomaly detection + root-cause classification |
| cnc-tool-wear-predictor | Tool wear state + RUL prediction — 96.8% accuracy, ~18-cut MAE |
| weld-quality-classifier | 5-class laser weld quality prediction with SHAP explainability |
| weld-signal-classifier | Real-time defect detection from 10 kHz photodiode + acoustic signals |
| laser-parameter-optimizer | Nd:YAG weld parameter recommendation for 7 aerospace alloys |
| milling-weld-quality-predictor | Two-stage cascade: milling surface quality → weld quality |
| additive-process-monitor | Layer-by-layer FDM printer failure detection — ~90% accuracy |
| anomaly-detection-dashboard | Multi-algorithm industrial sensor anomaly detection |
| rag-manufacturing-qa | RAG over manufacturing SOPs and maintenance docs |
| as9100-quality-tracker | AS9100 shop floor dashboard — FPY, OTD, NCR traceability |
| customer-growth-forecasting | ARIMA vs Prophet vs XGBoost holdout comparison — Prophet best (MAE 1,781) |
Python · XGBoost · scikit-learn · PyTorch · SHAP · LangChain · ChromaDB · Streamlit · SQL · Spark · Airflow · Docker · AWS
M.S. Data Science, UCLA (in progress) · Data Science Certificate, UCLA Extension · M.S. Engineering (Microelectronics), MIEM
etozhejimmy@gmail.com · LinkedIn · Kaggle · Northridge, CA