InspectorAI Pro
Team led by a Woxsen University researcher and Guinness World Record achiever skilled in Python, GenAI, CNNs, and deploying scalable ML pipelines on Google Cloud.
Project Description
The Agentic Engineering Tutor is designed to bridge the gap between complex industrial data and actionable human knowledge by replacing traditional, static dashboards with a generative interface.
Primary Users
Engineering Students & Trainees: Those learning to identify material failures using datasets like NEU-CLS can use the “Beginner Mode” to receive guided, visual walkthroughs of specific defects like Crazing or Inclusion.
Quality Assurance (QA) Inspectors: Factory floor personnel who need immediate, AI-driven second opinions on surface defects during the manufacturing process.
Senior Metallurgical Engineers: Experts who require “Data-Dense” views, such as pixel intensity histograms and baseline comparisons, to make high-stakes “Accept/Reject” decisions.
Key Benefits
Eliminates “SaaS Bloat”: By using the “Kill the Dashboard” approach, users aren’t overwhelmed by unnecessary menus; the UI only generates the specific tools or charts needed for the current task.
Accelerates Learning: The A2UI (Agent-to-UI) protocol transforms a standard chat into a “Live Lab,” allowing students to interact with 3D models or heatmaps rather than just reading text.
High-Precision Analysis: By leveraging Google Vertex AI (Gemini 1.5 Pro), the system provides structured JSON data (bounding boxes and confidence scores) that ensures industrial-grade accuracy.
Scalable & Cost-Effective: Utilizing Google Cloud credits and serverless architecture (GCP Cloud Functions) allows the system to scale without the overhead of maintaining local heavy infrastructure like Docker.
Prior Work
N/A