Rambabu Arabandi

AI Engineer · Hyderabad, India

Open to AI Engineer roles · Voice AI, agents & developer tooling

Rambabu Arabandi

I build AI that works in production: voice agents handling 500+ daily calls, MCP developer tools, and enterprise GraphRAG workflows evaluated with RAGAS and DeepEval.

Agentic systemsMCP toolingGraphRAG & evaluationVoice AIFull-stack
500+
daily calls on a production voice agent
End to end
from API integration to AI evaluation

Experience

  • Built an internal Go CLI that generates runnable MCP server projects, including boilerplate, tool and resource registration, and configuration wiring; tested the tooling with DeepEval
  • Contributed to an internal GraphRAG pipeline combining vector similarity search and graph traversal for related-document queries; evaluated the pipeline with RAGAS and DeepEval
  • Built and deployed n8n automation systems: multi-step data pipelines with webhook-driven agent triggers and event-based orchestration that replaced manual handoffs and ran unattended in production
  • Architected and shipped Zaya, a production real-time voice AI agent handling 500+ daily dealership calls, with a full STT to LLM tool-calling to TTS pipeline on LiveKit Agents with real-time turn management
  • Built multi-tool agentic workflows with dynamic routing, memory-driven conversations, and structured tool contracts
  • Owned the DMS integration layer: auth flows, webhook handling, and structured data exchange with dealership APIs
  • Shipped full-stack features across billing, CRM, and inventory on a live POS system (React, Node.js/Express)
  • Reduced production defects through systematic edge-case handling and improved test coverage
  • Worked across the full MERN stack (MongoDB, Express, React, Node.js) on the FitPaa website, covering both frontend and backend

Featured projects

Production voice AI, enterprise retrieval, and developer tooling. Explore the problem, my contribution, and the outcome behind each project.

Skills

AI & Agents

LLM IntegrationsPrompt EngineeringTool CallingAgent OrchestrationStructured OutputsGuardrailsMulti-Agent Systems

MCP, Tooling & Automation

MCP ServersCLI Toolingn8n WorkflowsWebhook-Driven PipelinesEvent-Based OrchestrationAPI Integration

Voice AI

LiveKit AgentsSarvam AIReal-time STT/TTSLow-Latency VoiceTurn Management

RAG & Retrieval

Hybrid RetrievalGraphRAGBM25Vector SearchGraph RetrievalFalkorDBRRF FusionCross-Encoder Reranking

AI Evaluation

RAGASDeepEvalRAG Pipeline EvaluationMCP Tooling Evaluation

AI Frameworks

LangChainLangGraphVercel AI SDKClaude CodeCursor

Full-Stack

FastAPINode.js/ExpressREST APIsWebhooksNext.jsReactTailwind CSS

Data, Cloud & DevOps

MongoDBPostgreSQLSupabaseAWSGCPDockerVercelGit

Languages

PythonTypeScriptJavaScriptGo

Speaking & Education

2026

Speaking & Teaching

Speaker at NOVA 2026 (ACM GRIET) on MCP, A2A protocols, and AI-native development. Teaches AI engineering (agentic systems, RAG, and production deployment) to students, professionals, and founders.

2019 - 2023

B.Tech, Computer Science Engineering

Bharat Institute of Engineering and Technology, Hyderabad

Beyond code

Away from the keyboard: music, and taking hardware apart to see how it works. Move your pointer over each figure.

Singing

I sing for the love of it. Music is how I switch off after a long day of debugging.

  • Vocals
  • Melody
  • Live jams

Hardware & repairs

I learn what's inside phones and laptops: glass, boards, batteries, fans and shells. I open them up and take on repairs now and then.

  • Teardowns
  • Repairs
  • Upgrades

Custom ROMs & recovery

I love giving older Android phones a second life: unlocking the bootloader, flashing a custom recovery and a newer ROM, and rooting.

  • Custom ROMs
  • Recovery
  • Rooting

About

I'm Rambabu Arabandi, an AI Engineer in Hyderabad, India. I build AI that works in production: voice agents, GraphRAG retrieval workflows and MCP developer tooling, on a full-stack background in Next.js, TypeScript, Python (FastAPI) and Go.

Zaya, a voice agent I built, answers 500+ calls a day for a US auto dealership. I also built a GraphRAG pipeline that combines vector search with graph traversal, evaluated with RAGAS and DeepEval, and a Go CLI that scaffolds runnable MCP servers in one command.

I've worked at Arohak (on a Johnson & Johnson engagement via Cognizant), Cyepro Solutions, Ordermatic Technologies and Gymclan Fitness Services. I spoke at NOVA 2026 (ACM GRIET) on MCP and A2A protocols, and I'm open to AI Engineer roles.

Contact

Open to AI Engineer roles. Let's talk.

Based in Hyderabad, India. The fastest way to reach me is email, and I'm also active on LinkedIn and GitHub.

rambabuarabandi2001@gmail.com