Software Engineering Senior Engineer II - Agent Harness & Workflow Engineering Professional Mul

At IBM Software, we transform client challenges into solutions. Building the world’s leading AI-powered, cloud-native products that shape the future of business and society. Our legacy of innovation creates endless opportunities for IBMers to learn, grow, and make an impact on a global scale. Working in Software means joining a team fueled by curiosity and collaboration. You’ll work with diverse technologies, partners, and industries to design, develop, and deliver solutions that power digital transformation.

With a culture that values innovation, growth, and continuous learning, IBM Software places you at the heart of IBM’s product and technology landscape. Here, you’ll have the tools and opportunities to advance your career while creating software that changes the world.

Your role and responsibilities

We are an agent-first engineering team building intelligent infrastructure automation  platform combines frontier and small language models with strong systems engineering to deliver production‑grade automation backed by rigorous evaluation, telemetry, and quality gates. You’ll work side‑by‑side with both humans and AI agents, where writing code and orchestrating agents are equally core to the role.

What You’ll Do

• Build and evolve type-safe, composable AI pipelines for intent analysis, planning, knowledge enrichment, and infrastructure generation
• Extend pipeline routing and outputs across Z, Power, and Cloud platforms (Terraform, Ansible, Pulumi, native APIs)
• Design context management, telemetry, and production observability for multi-step AI workflows
• Own a markdown-based workflow engine supporting parallel execution, rollback, approvals, and cross-domain orchestration
• Develop and scale a plugin-based tool provider system
  , integrating Power and Cloud APIs with secure session management
• Advance a robust evaluation framework (LLM-as-judge, regression detection, eval-as-CI) to ensure AI-generated outputs are production-safe
• Help build a cross-platform workflow catalog used by humans and AI agents for discovery, routing, and execution
• Extend the agent-to-agent (A2A) interface for capability discovery and delegation as the platform scales

What the First 90 Days Look Like

Month 1: Onboard onto the Go agent codebase. Run the full eval suite, understand the pipeline composition model, execute scripted workflows against the infrastructure simulator. Trace a request end-to-end from the A2A interface through intent analysis, pipeline routing, tool dispatch, and infrastructure code output. Understand the eval case format, run cross-model comparison, review regression baselines.

Month 2: Extend the eval corpus — add new cases covering under-tested pipelines (iterative planning, deep research, workflow composition). Begin workflow catalog design — metadata schema, platform tagging, versioning model. Prototype a Power tool provider interface based on Power HMC API surface.

Month 3: Ship workflow catalog MVP with the existing Z workflows cataloged and tagged. Deliver first Power or Cloud tool provider prototype. Implement eval-as-CI with quality gates. Begin cross-platform pipeline routing — intent analysis that distinguishes Z, Power, and Cloud prompts and routes to the correct pipeline/tool provider combination.

Required education

High School Diploma/GED

Preferred education

Bachelor's Degree

Required technical and professional expertise

· Strong Go engineering. The entire harness is Go — pipeline composition, workflow execution, tool dispatch, eval framework, agent protocol server. You need to be productive in Go immediately.

· You’ve built execution systems. Workflow engines, pipeline frameworks, DAG executors, orchestration infrastructure — you understand step composition, error propagation, checkpoint/rollback, and what it takes to make multi-step execution reliable.

· LLM integration experience. You’ve worked with LLM APIs in production — prompt construction, structured output parsing, function/tool calling, context window management. You understand the practical challenges of making LLM outputs reliable.

· You care about quality measurement. You’ve built or maintained evaluation, testing, or quality systems. You…