Website

PRIME (Predictive Reliability & Intelligence Maintenance Engine)

An AI-driven machine condition monitoring platform integrating hybrid ML/DL predictive models, a RAG-based technical assistant, and automated maintenance scheduling within a unified industrial ecosystem.

Role

ML Engineer & AI Integrator

Year

2026

Client

Project-Based Learning

Technologies

Next.jsFastAPIPythonTensorFlowXGBoostRAGDockerSocket.IO
Project Main Visual

The Challenge

Manufacturing facilities relied heavily on corrective or time-based maintenance approaches, resulting in sudden operational downtime and reducing productivity by up to 30%. Furthermore, existing predictive systems were isolated from operational workflows and factory knowledge bases, imposing significant cognitive overhead on technicians during data interpretation.

The Solution

Architected a centralized AI system using a monorepo structure to orchestrate two primary artificial intelligence integrations. First, engineered an ML Engine utilizing Python and FastAPI to deploy an XGBoost model as a Health Status Classifier and an LSTM network for cascading Remaining Useful Life (RUL) predictions. Second, developed a RAG-based NLP Engine functioning as an AI Copilot to resolve technical queries against factory SOPs. Model inferences were subsequently routed through a Node.js backend to generate automated rule-based maintenance schedules, broadcasted in real-time to a Next.js dashboard via Socket.IO.

Visual Showcase

Gallery image 1
Gallery image 2
Gallery image 3
Gallery image 4
Gallery image 5
Next Case Study

GAGASAN