Earlier, I built an application that detects PPE compliance hazards in the steel manufacturing industry and generates real-time logs.
While the detection worked well, the real challenge emerged afterward: huge volumes of logs that are extremely hard to manually review and convert into meaningful insights.
That's what led me to build AutoPPE-Analyst.
How it works
AutoPPE-Analyst is a Python + AutoGen multi-agent application that transforms PPE CSV logs into clear, compliance-ready reports using Gemini (via an OpenAI-compatible client).
- Multi-agent orchestration: Data Preparer → Analyst → Reporter
- Prompt-driven behavior contracts with structured Pydantic outputs
- Produces reports designed for non-technical stakeholders, not just raw metrics
- Supports both CLI batch runs and an interactive Streamlit UI
Why it matters
- Eliminates hours of manual log review
- Converts operational data into actionable compliance insights
- Bridges the gap between AI detection systems and decision-making