I had an interview with Fast Accounting for a machine learning role. The discussion focused on my experience with deep learning, OCR, and computer vision. They asked technical questions about model optimization and anomaly detection. I also presented recent work on defect detection in CT scans using PatchCore and RF-DETR. The team seemed interested in practical implementation and MLOps workflows for industrial AI solutions.
Interview questions [1]
Question 1
They asked me to explain how I combined unsupervised and supervised methods—like PatchCore and RF-DETR—for defect detection in battery CT scans, and how I ensured both accuracy and low false positives in real-world applications.