Arulvelu G.
Applied AI in Production

AI & Intelligent Systems

Production-grade AI systems designed to automate operations, reduce manual effort, and enable intelligent decision-making at scale.

Built using LLMs, RAG, APIs, and event-driven architectures
Featured System

Enterprise Knowledge Retrieval System (RAG)

LLMEmbeddingsVector DBLangChainOpenAI
Built a Retrieval-Augmented Generation (RAG) system to enable intelligent querying of enterprise documentation → reduced manual search effort and improved knowledge accessibility.
Documents → Embeddings → Vector DB → LLM Retrieval → Response
Faster knowledge retrieval · Reduced manual lookup effort
Featured System

AI-Driven Invoice Classification System

OCRAzure OpenAIAPIs
Automated invoice classification using OCR and LLMs to extract and categorize financial data → reduced manual processing effort and improved accuracy.
OCR → LLM → Classification → Output
Reduced manual effort · Improved accuracy

AI-Powered Conversational Workflow Assistant

LLMTool CallingAgents
Built a conversational AI assistant to trigger workflows and retrieve system data → reduced manual interactions and improved operational efficiency.
User Query → LLM → Tool Calling → Workflow Execution
Reduced manual interaction · Faster task execution

Enterprise Knowledge Automation with MCP

MCPJira APIConfluence
Developed an MCP-based system integrating Jira and Confluence to enable intelligent querying and workflow automation → improved knowledge access and reduced manual effort.
Query → MCP → Jira/Confluence → Response
Faster information retrieval · Reduced manual effort

AI-Assisted Automation Reporting Dashboard

AI CopilotData AnalyticsDashboards
Built a reporting dashboard using AI copilots to generate insights from automation data → enabled faster decision-making and reduced manual analysis.
Data → AI → Insights → Dashboard
Faster insights · Reduced reporting effort

Intelligent Lead Qualification System

AI ClassificationLead Routing
Developed an AI-driven system to classify and prioritize leads → improved response time and conversion efficiency.
Input → AI Classification → Routing
Improved response time · Increased conversions

AI-Based Incident Triage & Routing System

LLMITSMTriage Classification
Designed an AI-powered system to classify incidents and route them automatically → reduced manual triage effort and improved resolution speed.
Ticket → LLM → Classification → Routing
Reduced triage time · Faster resolution