MachInsight-AI---Machine-Sensing-and-Monitoring-AI-Agent

Application URL : https://machinsight-ai-297752632164.us-central1.run.app

**Sample Prompts : **

1. plot the temperature trend of machine 1

2. are there any machines with temperature greater than 45 in last one hour

3. When was the last time the machine 1 turned faulty

4. How many machines have turned faulty in last 2 hours?

   

⚙️ MachInsight AI – Real-Time Machine Sensing & Monitoring AI Agent

MachInsight AI is an intelligent real-time platform that empowers manufacturers to monitor sensor data, predict faults, and interact with their data using natural language — all powered by AI, machine learning, and vector search.

---

🚀 Inspiration

In today's fast-paced manufacturing environment, machine downtime can lead to massive productivity loss. We were inspired to build MachInsight AI to empower industries with a real-time, intelligent monitoring system that not only detects potential faults in advance but also enables users to interact with their sensor data through natural language.

Our goal: bring together the best of data warehousing, machine learning, vector databases, and AI agents into a unified solution.

---

🤖 What It Does

MachInsight AI is a real-time IoT monitoring and diagnostics platform powered by AI. It:

• Ingests real-time sensor data into BigQuery
• Aggregates data hourly and predicts fault probability via Random Forest ML
• Stores results and semantic embeddings in MongoDB
• Builds a vector index with MongoDB’s vector search
• Powers an AI Agent using Google ADK + Gemini 2.0 Flash that can answer:
  • “Are there any machines with temperature greater than 45 today?”
  • “What’s the last time Machine 2 turned faulty?”
  • “Find machines similar to Machine 5’s fault behavior.”

---

🧱 How We Built It

• Data Ingestion: Real-time IoT sensor data sent to BigQuery
• ML Pipeline: Vertex AI Colab notebooks compute hourly summaries and apply a trained Random Forest model
• Storage & Embeddings: Output stored in MongoDB; embeddings created and indexed for vector search
• Agent & Tools: Agent built with Google ADK, including tools like:
  • machines_with_high_metrics
  • last_faulty_time
  • faulty_machines_summary
  • plot_machine_trend
  • find_similar_machine_events
• Scheduling: GCS and cron jobs orchestrate the notebook execution

---

🧗 Challenges We Ran Into

• ⏱ Notebook Scheduling: Ensuring ML and embeddings pipelines run in the right sequence
• 🧠 Natural Language Parsing: Mapping vague user questions to precise MongoDB queries
• 🎯 Vector Search Tuning: Crafting meaningful embeddings and similarity thresholds
• 🔐 Access Issues: Solving signed URL and GCS permission problems during dynamic chart generation

---

🏆 Accomplishments We're Proud Of

• A seamless end-to-end pipeline from sensor ingestion to AI interaction
• Modular agent with plug-and-play tools for analytics and visualization
• Harmonized stack: BigQuery, Vertex AI, MongoDB, ADK, Gemini
• Real-time, intelligent interaction on IoT data with predictive maintenance insights

---

📚 What We Learned

• ML pipeline orchestration using Vertex AI Colab Enterprise
• Semantic search with MongoDB vector search
• How to build AI agents with Google ADK + Gemini, including multi-turn contextual chat
• Best practices for data modeling and real-time analytics in IoT

---

🔮 What’s Next for MachInsight AI

• Integrate anomaly detection using autoencoders or LSTM models
• Expand support for image/video sensors for visual fault detection
• Enable automated triggers for repairs and alerting
• Build a Looker Studio dashboard + embed agent chat widget
• Explore Edge AI deployment to run insights at the factory floor

---

🧰 Tech Stack

• BigQuery – Real-time data warehouse
• Vertex AI – ML model training and prediction pipelines
• MongoDB + Atlas Vector Search – Data & embeddings storage with semantic search
• Google ADK + Gemini – AI Agent platform with LLM integration
• Cloud Run / Docker – Deployment infrastructure

---