Senior Agentic Harness Engineer

We are looking for a Senior Agentic Harness Engineer to build and operate a production-grade, AWS-native agentic harness for autonomous software delivery.

Responsibilities:

• Build and operate agentic loops covering orchestration, agent execution, coding, review, testing, and JIRA integration.
• Implement LLM orchestration, model routing, retry logic, and provider fallback mechanisms.
• Manage token costs, context windows, agent budgets, and circuit breakers.
• Build observability, monitoring, and alerting for agent health, cost, and quality.
• Implement agent governance, safety guardrails, kill switches, audit logging, and prompt-injection mitigation.
• Integrate JIRA/Atlassian APIs and other business systems.
• Work closely with client engineering teams on architecture, development, demos, and capability transfer.

Requirements:

• Experience: 3-5+ years, with meaningful experience in agentic AI/LLM orchestration.

Must-Have Skills:

• Production experience building Agentic AI systems.
• LangChain, LangGraph, or equivalent agent orchestration frameworks.
• LLM orchestration and model routing across multiple providers/models.
• Agent governance, human-in-the-loop workflows, guardrails, and audit trails.
• Prompt injection prevention and mitigation.
• Token cost management and context window/memory handling.
• Strong Python development experience.
• AWS experience; Bedrock/AgentCore exposure is a strong plus.
• Observability/monitoring tools such as Langfuse, Datadog, or equivalent.
• Experience with JIRA/Atlassian APIs or similar ticketing systems.

Nice to Have:

• AWS Bedrock AgentCore.
• Temporal or similar workflow orchestration.
• LiteLLM.
• Cursor or other AI-native IDEs.
• Self-hosted open-weight models such as DeepSeek/GLM.
• Claude Code, Claude Cowork, or Claude Skills.
• Financial services or regulated-industry experience.