Looking for a job

Ragul Manthiramoorthy

@ragulmanthiramoorthy

Message

Agentic AI Systems Engineer at Rently building LangGraph voice agents and an in-house MCP server.

India

Message

At Rently, I developed an agentic AI application using LangChain and LangGraph, including LiveKit-powered voice agents for conversational workflows. I also built an in-house MCP server so agents can execute rental application workflows through structured tool calls.

As Associate Software Engineer, I owned Kafka data pipelines and implemented PostgreSQL CDC and JDBC sink pipelines for event-driven services. Migrating self-managed Kafka to Kubernetes yielded approximately $60K in annual savings.

On my personal AI-Powered Travel & Fleet Management Platform, I designed agent workflows that connect live actions and policy knowledge through MCP and RAG. I also built voice features and an evaluation layer for retrieval quality, faithfulness, tool selection, and hallucination detection.

Experience

Work history, roles, and key accomplishments

Current

Software Engineer

Current

Rently

Apr 2026 - Present (6 months)

Developed an agentic AI application using LangChain with LangGraph orchestration, including LiveKit-powered voice agents for conversational workflows. Built an in-house MCP server for the rental-homes web application, exposing business capabilities as tools so AI agents execute application workflows through structured tool calls.

Lang Chain LangGraph MCP LiveKit AI Agents Voice AI SMS Email Python fastAPI

Associate Software Engineer

Rently

Apr 2024 - Apr 2026 (2 years)

Migrated and operated self-managed Kafka 3.8 on Kubernetes (Strimzi, KRaft) in place of Confluent Cloud, yielding approximately $60K in annual savings. Implemented Debezium PostgreSQL CDC and JDBC sink pipelines powering event-driven services, and developed an OAuth 2.0 authorization server in Ruby on Rails.

Kafka Kubernetes Strimzi Debezium PostgreSQL Redis Ruby on Rails OAuth Google Geolocation API CI CD RSpec

Junior Software Engineer

Rently

Apr 2023 - Apr 2024 (1 year)

Implemented Rails-based business rules and Kafka event-driven processing for the Trespasser Alert System, handling events from IoT devices such as locks, sensors and thermostats. Integrated Persona and Sift for renter fraud prevention, then developed an in-house scoring approach that replaced the external services, saving approximately $2,000/month.

Ruby on Rails Kafka IoT Personas SIFT Fraud Prevention Event Driven Architecture