AI Agent / GenAI Developer
Shankarshan MishraUnited StatesMIDPosted 45m ago
PythonLLM/AI AgentsLangGraphLangChainRAGVector SearchCOBOLJCLDB2FastAPIReactAngularAzure OpenAIOpenAIBedrockNeo4jGraphRAGPurviewCollibra
About the Role
We are looking for a hands-on AI Agent / GenAI Developer to build an AI-powered platform that analyzes legacy mainframe applications and generates end-to-end technical and attribute-level data lineage. You will build AI agents that parse COBOL, JCL, DB2 and Copybooks, extract dependencies, and deliver a web interface where users can search any attribute, table, file or program and explore its full upstream and downstream lineage.
Key Responsibilities
• Build AI agents and GenAI workflows to analyze mainframe source code and technical artifacts
• Analyze COBOL, JCL, DB2 SQL, Copybooks, VSAM/files and batch flows to identify dependencies
• Extract metadata: programs, jobs, files, tables, columns, transformations and field-level relationships
• Generate end-to-end lineage: Attribute → Copybook → Program → DB2/File → JCL/Job → Transformation → Target
• Implement RAG, embeddings, vector/hybrid search and agentic workflows over large code repositories
• Build knowledge-graph models for navigating application and data dependencies
• Develop backend services and REST APIs using Python and FastAPI (or equivalent)
• Build or integrate a React/Angular web UI for lineage search and visualization
• Enable natural-language queries such as: Where does this attribute originate? Which programs modify this field? What is the complete upstream/downstream lineage? Which applications are impacted if this attribute changes?
• Ensure AI-generated lineage is grounded in, and traceable to, actual source-code evidence
Requirements
• Must have: Python + LLM/AI Agents (LangGraph/LangChain) + RAG/Vector Search + COBOL/JCL/DB2 knowledge + FastAPI + React/Angular + Azure OpenAI/OpenAI/Bedrock
• Bonus: Neo4j/GraphRAG, Purview/Collibra, mainframe modernization