Academic
AI-Powered Industrial Monitoring (TÜBİTAK 3501)
An end-to-end OCR & NLP based monitoring system analyzing production-line inefficiencies. Part of my M.Sc. thesis and a TÜBİTAK 3501 project proposal.
Academic · OCR + NLP · ongoing
PythonOpenCVTrOCRTesseractBERTPandas
Problem
On industrial production lines, downtime reasons were recorded on handwritten forms and entered as free text; the data could not be analyzed and root-cause analysis was impossible.
Solution
As a research engineer at Sakarya University, I am building an AI-based monitoring system that digitizes handwritten production forms and standardizes operator inputs.
Highlights
- Image-processing pipelines with Python + OpenCV
- A hybrid OCR architecture combining TrOCR and Tesseract
- BERT-based NLP models standardizing free-text downtime reasons
- Automated data cleaning/validation pipelines with Pandas
- Contribution to a decision-support system for production analytics