Principal Harness Engineer

This is a fully remote position, open to applicants in Bulgaria.

📋 Description

• Design, construct, deploy, and maintain core harness elements, including guides, feedback loops, guardrails, and shared context.

• Take ownership of and enhance feedforward guides, such as agent instruction files, reusable skills, architectural guidelines, reference documents, and codemods.

• Promote team-wide standardization for outputs generated by agents.

• Develop feedback sensors, including custom linters, static analysis tools, structural and architecture-fitness tests, verification loops, and LLM-as-judge reviewers.

• Manage quality gating and establish release criteria for outputs produced by agents.

• Define authority boundaries for unaided agent merges and create escalation protocols for human review.

• Create LLM testing frameworks and assessment strategies for the quality and safety of AI-generated outputs.

• Implement consumer/contract testing when service integration reliability is crucial.

• Operate the steering loop to avoid recurrent agent errors and maintain repository knowledge as the system of record.

• Determine the running of controls across pre-commit, post-integration, and ongoing drift-monitoring phases.

• Enhance visibility into agent work and monitor metrics such as cost per merged PR, time-to-merge, review velocity, defect escape rate, and agent-PR survival rate.

• Collaborate with product and platform teams to translate specifications and acceptance criteria into enforceable controls.

• Act as a technical expert and mentor engineers on harness practices and responsible use of AI tools.

⛳️ Requirements

• Bachelor's or Master's degree in Computer Science or related fields, or equivalent commercial experience in software architecture, design, development, and testing.

• Proven experience as a seasoned software engineer with extensive, complex codebases.

• Strong grounding in architecture and design principles.

• Practical experience with AI coding agents such as Claude Code, Codex, or similar technologies.

• Experience in developing engineering tools within a modern stack, including linters, static analysis, CI/CD pipelines, containerized build/test environments, and instrumentation/observability.

• Familiarity with agent instruction conventions, such as AGENTS.md.

• Experience with specification-driven development, context engineering, agent orchestration, fitness functions, and developer-platform collaboration.

• Capability to encode engineering standards into mechanical, repeatable rules.

• Judicious selection of deterministic computational controls versus inferential LLM-based controls, with insight into cost, speed, and reliability trade-offs.

• Experience in managing quality-gating processes and establishing release criteria.

• Working knowledge of autonomous-agent security, including prompt injection, tool/permission scoping, sandboxed execution, audit trails, and least-privilege guardrails.

• Familiarity with consumer/contract testing methodologies, such as Pact.

• Exceptional communication skills, with the ability to articulate designs with architects and promote standards across teams.

🏝️ Benefits

• Highly competitive compensation, benefits, and rewards programs.

• Focus on work/life balance.

• Employee resource groups.

• Social events.

• Opportunities for professional development and valuable learning experiences.

• An inclusive, people-first work environment.

• Equal employment and advancement opportunities regardless of protected characteristics.