Ajay

Hi, I am Ajay Balu Kumar Mithireddy, a Generative Al and Agentic Al Engineer based in Hyderabad.

Currently I am working at CloudTern, where I build production-ready Al systems. My main focus is on Agentic RAG and multi-agent workflows. I build systems where multiple Al agents collaborate, use tools, and interact with enterprise data to solve real business problems.

I work extensively with Python, FastAPI, LangChain, LangGraph, CrewAl and AutoGen, along with LLMs from OpenAI and Azure OpenAI, and AWS Bedrock. On the data side, work with vector databases like Pinecone and Weaviate for RAG pipelines, embeddings and retrieval. For deployment, I use Docker, Kubernetes, AWS and Azure.

What drives me is moving Al from a proof-of-concept to a governed, reliable production system. I enjoy owning the end-to-end flow - from ingestion and chunking, to retrieval and re-ranking, to agent orchestration, evaluation, and monitoring in production.

I am Ajay Balu Kumar Mithireddy, a Generative AI and Agentic AI Engineer based in Hyderabad. I am currently working as a Software Engineer - Generative AI and Agentic AI at CloudTern Solutions from Feb 2025 to Present.

Experience at CloudTern:

1. Ansera - Agentic RAG Search Platform: I architected an agentic RAG search engine using LangChain and LangGraph that turns client websites into conversational interfaces. I designed hybrid dense and sparse retrieval with a re-ranking layer, and built multi-tenant, multi-source ingestion pipelines on Weaviate with incremental updates, audio/video transcription, and OCR-powered extraction.

I also implemented governance guardrails to improve answer faithfulness, built an AEO and GEO scoring tool with MCP and A2A protocols, and automated site onboarding via Webflow and WordPress API integrations.

2. MCP Server Development: I engineered an authenticated, role-based MCP server for natural-language access to website analytics across client sites, and built Calendar and Calendly MCP tool integrations for lead capture and meeting scheduling inside chat.

3. PolisIQ - AI-Powered Insurance Document Intelligence: I built LangGraph agent workflows and MinerU OCR extraction pipelines that convert scanned insurance documents into structured data, and deployed Gemma model with vLLM on Dockerized RunPod GPU workers for production-grade inference.

4. Resume Tracker - AI Resume Search Engine: Built a LangGraph and Neo4j based resume search using a graph model of skills, experience and education, with LLM-generated Cypher queries for fast candidate discovery.

Before CloudTern, I was an Azure Cloud and DevOps Engineer Intern at Revature from Apr 2024 to Aug 2024, where I deployed a full-stack application on Azure with GitHub CI/CD, and worked with AKS, Terraform, Prometheus, Grafana and MySQL.

Certifications and Achievements:

• Microsoft Certified: Azure AI Apps and Agents Developer Associate (AI-103), 2026
• GitHub Certified: Agentic AI Developer (GH-600), 2026
• *Oracle Certified: Agentic AI Certified Foundations Associate (1Z0-1157-26), 2026

My core skills include LLMs, RAG, Agentic AI, Multi-Agent Systems, Tool Calling, LangChain, LangGraph, LlamaIndex, CrewAI, AutoGen, MCP, A2A, FastAPI, Pinecone, Milvus, Weaviate, ChromaDB, Neo4j, Azure AI Foundry, AWS Bedrock, and Document Intelligence with MinerU and PaddleOCR.

B.Tech in Computer Science and Engineering from Pragati Engineering College, Kakinada, 2019-2023