Rahul Korde

AI Software Engineer

Talent

PuneMember since July 12, 2026

PythonSQLAgentic AIRetrieval-Augmented Generation (RAG)Generative AI (GenAI)Large Language Models (LLMs)Multi-Agent SystemsLangChainLangGraphAgent-to-Agent (A2A) ProtocolMicrosoft AzureAzure OpenAIExtract, Transform, Load (ETL) DevelopmentDocker

Hire this person

Send a job offer directly to this candidate

PythonSQLAgentic AIRetrieval-Augmented Generation (RAG)Generative AI (GenAI)

About

AI Software Engineer with 4.9+ years of experience developing Agentic AI and Retrieval-Augmented Generation (RAG) systems using Large Language Models (LLMs). Delivered production multi-agent platforms on LangChain, LangGraph, Agent to Agent (A2A) and Azure OpenAI, reducing manual engineering effort by up to 85% and cutting full-suite test turnaround time by 70%. Skilled in Python, SQL, Extract, Transform, Load (ETL) development, and enterprise data engineering on Microsoft Azure, applying agentic-architecture and Generative AI (GenAI) expertise to convert AI prototypes into measurable business impact.

Experience

AI SW Engineer at Amdocs (2021-08 – Present)

Build agentic AI and GenAI products (AOS, DeepQE) using Python automation and RAG architectures to support enterprise-scale AI workflows.

Built an AI-powered test-case automation platform for ETL/Data Warehouse pipelines, cutting manual SQL test-case authoring effort by 85% and accelerating data-migration go-live timelines.
Delivered a 7-category multi-agent validation system (uniqueness, integrity, completeness, reconciliation, schema, SCD Type 2), achieving end-to-end test coverage across 100+ tables per run and cutting full test-suite turnaround time by 70%.
Built a feedback-driven self-correction loop that learns from reviewer input, raising first-pass SQL accuracy to 92% and cutting repeat-defect rework by 60% — directly lowering QA cost across data-warehouse deliveries.
Built an A2A protocol microservice that automates COBOL copybook-to-VSAM mapping, cutting manual reverse-engineering effort by 85% and accelerating mainframe migration go-live.
Architected a 10-step ingestion-to-inference pipeline (COBOL/copybook/JCL parsing, graph-based evidence linking, sample-record matching) orchestrated by a 6-agent Agno team on Azure OpenAI, eliminating manual copybook-matching effort at scale.
Shipped a production-ready, Docker-containerized streaming A2A skill generating audit-ready mapping reports and orphan indexes, validated by a 54+ test suite before release.