
About this role
The candidate will design, build, and improve an AI-driven migration factory that uses agentic automation to convert legacy COBOL systems into modern software. Responsibilities include developing Python-based services, tools, and execution pipelines to support the end-to-end migration process. The role involves configuring and extending AI agents using Claude Code and MCP, and implementing orchestration logic for multi-step workflows. The engineer will work closely with COBOL experts and technical stakeholders to refine prompts, agent skills, and validation logic based on feedback. The technical environment focuses on autonomous development workflows and multi-agent systems.
Skills & technologies
Must have
- Python
- Agentic AI
- LLM
- Multi-agent systems
- Claude Code
- MCP
- AI-assisted software engineering
Nice to have
- RAG
- Vector databases
- Repository indexing
- Codebase knowledge systems
- LangGraph
- LangChain
- CrewAI
- OpenAI Agents SDK
Read full description
In this role, you will help design, build, and continuously improve an AI-driven migration factory that converts legacy COBOL systems into modern software through agentic automation. Your focus will be on engineering reliable automation that analyzes, generates, validates, and continuously enhances migration outcomes rather than manually rewriting legacy applications.
Responsibilities
- Design, develop, and maintain an AI-powered migration factory for automated legacy system modernization
- Build agentic workflows that analyze COBOL applications, generate modern code, validate outputs, and support iterative refinement
- Develop Python-based services, tools, integrations, and execution pipelines that support the end-to-end migration process
- Configure and extend AI agents using Claude Code, MCP, structured prompts, tools, skills, and project-specific context
- Implement orchestration logic for multi-step and multi-agent modernization workflows
- Integrate agents with source code repositories, documentation, build systems, testing frameworks, and other engineering tools
- Review generated outputs together with the client’s COBOL experts and technical stakeholders
- Translate expert feedback, identified defects, and business-rule gaps into improvements to prompts, agent skills, validation logic, and workflow design
- 5+ years of hands-on experience in Python software engineering
- Proven experience designing and implementing Agentic AI, LLM, or multi-agent systems
- Hands-on experience with AI-assisted software engineering and autonomous development workflows
- Experience with Claude Code or comparable agentic coding environments
- Experience integrating external tools and data sources through MCP, APIs, command-line tools, or custom adapters
- Upper-intermediate or higher English level
- Familiarity with RAG, vector databases, repository indexing, or codebase knowledge systems (as a plus)
- Experience with LangGraph, LangChain, CrewAI, OpenAI Agents SDK, or similar orchestration frameworks (as an advantage)