Manufacturing·AI & Machine Learning + Business Automation
Harlow Manufacturing
A precision parts manufacturer cut defect escape rate by automating visual inspection on the line.
100%
of parts inspected, up from ~15%
78%
reduction in defect escape rate
< 400ms
inspection time per part
The challenge
Harlow's quality team hand-inspected a sample of parts per batch — enough to catch major defects, not enough to catch the subtle ones. Defective parts occasionally reached customers, triggering costly returns and strained relationships.
The solution
We deployed a computer vision inspection system at the end of the line, trained on Harlow's own defect history, that inspects every part instead of a sample. Flagged parts route to a human reviewer with the specific defect highlighted, keeping a person in the loop on every rejection.
Technology stack
PythonPyTorchEdge inference (NVIDIA Jetson)Custom dashboard
“We went from sampling to full inspection without adding headcount. That's the part our customers actually feel.”
Let's talk about what's slowing you down.
Book a 30-minute call. We'll tell you honestly whether we're the right fit before we talk about scope.