AI-powered electrical diagnostics from images.
PythonClaude VisionFastAPISupabaseRender
Problem
Electrical inspections are slow, manual, and dependent on scarce expert knowledge. Scaling expertise is expensive and error-prone.
Solution
A system that analyzes images of electrical panels, schematics, and installations to detect faults, assess risks, and recommend actions in seconds.
How AI is used
Built on top of Claude Vision. Instead of training models, I designed a structured prompt system that embeds expert knowledge directly into the AI context — enabling consistent, high-quality analysis without training data.
Results
Reduced analysis time from hours to minutes. Increased consistency across diagnostics. Enables non-experts to perform expert-level assessments.