AgentX - AI-Powered Predictive Maintenance Platform
About
An AI-driven predictive maintenance platform for automobile manufacturing industries that monitors robotic arm servo motors, detects anomalies, automates incident workflows, and predicts maintenance cycles in real time.
Overview
Developed AgentX, an AI-powered predictive maintenance platform designed for the automobile manufacturing industry. The system continuously monitors robotic arm servo motors in production environments, analyzes telemetry data in real time, detects anomalies, predicts maintenance cycles, and automates incident management workflows.
The platform combines AI, cloud infrastructure, real-time streaming, and industrial automation concepts to improve operational reliability, reduce downtime, and optimize resource utilization.
Objectives
- Monitor robotic arm servo motors in real time
- Detect anomalies before hardware failure occurs
- Predict maintenance schedules based on resource usage
- Automate incident management and recovery workflows
- Analyze energy consumption and carbon emissions
Tech Stack
- Frontend: Next.js (PWA-enabled)
- Backend: NestJS
- Database: NeonDB + Prisma ORM
- Streaming & Messaging: Apache Kafka
- Realtime Data Sync: Firebase
- AI & Cloud Services: Azure AI, AWS
- Infrastructure: Scalable cloud-native architecture
Key Features
- Real-time telemetry monitoring from industrial sensors
- AI-based anomaly detection using servo motor profiles and operational capacity
- SOP-driven incident analysis and automated alert generation
- ServiceNow ticket creation for approval-based corrective actions
- Predictive maintenance engine based on historical usage patterns
- Carbon footprint analysis for each servo motor
- 3D simulation of robotic arm joints with energy consumption visualization
- Progressive Web App (PWA) support for cross-platform accessibility
Architecture / Design
Designed an event-driven, cloud-native architecture where telemetry data from servo motor sensors is streamed through Firebase and Kafka into the monitoring system.
The AI layer evaluates:
- Motor type
- Operational load
- Usage patterns
- Historical anomalies
- Standard Operating Procedures (SOPs)
When anomalies are detected:
- AI reviews the issue
- Generates operational alerts
- Creates ServiceNow tickets automatically
- Waits for human approval before implementing corrective actions
The architecture supports horizontal scalability and parallel processing for handling multiple robotic systems simultaneously.
Implementation
- Built real-time telemetry ingestion pipelines using Kafka and Firebase
- Implemented AI-based anomaly detection workflows
- Integrated SOP-driven incident evaluation logic
- Developed predictive maintenance algorithms using usage and performance trends
- Created automated ServiceNow ticketing workflows
- Designed interactive 3D visualization for robotic arm energy analytics
- Implemented carbon emission tracking per motor and joint
Outcomes
- Reduced unplanned downtime through proactive anomaly detection
- Automated incident workflows and maintenance recommendations
- Improved operational visibility with live telemetry dashboards
- Enabled predictive maintenance instead of reactive repairs
- Delivered energy and carbon footprint analytics for sustainability tracking
Impact
AgentX demonstrates how AI, cloud infrastructure, and industrial automation can transform manufacturing operations into intelligent, self-optimizing systems.
The platform combines:
- Predictive Maintenance
- AI-driven Decision Making
- Industrial IoT
- Sustainability Analytics
- Automated Incident Management
This aligns with modern Industry 4.0 and smart manufacturing practices.
Scalability & Reliability
- Event-driven architecture using Kafka
- Cloud-native deployment strategy
- PWA-enabled frontend for accessibility
- Designed for multi-factory scalability and real-time parallel monitoring
Role
- Full Stack / Cloud / AI Engineer