AI Harness Engineers

Location: Atlanta, United States

Salary: $150,000–$170,000 annually, plus potential variable compensation.

Company

An AI-native product organization within Capgemini Financial Services building software products for insurance claims, payment operations, and health operations.

What you will do

• Own fast, isolated, and reproducible development environments for human engineers and AI agent fleets, including sandboxes, ephemeral environments, and warm starts.
• Design deterministic CI and pre-push validation workflows that detect failures before code reaches the pipeline.
• Build and continuously improve the agent harness, including tooling, permissions, retry limits, and orchestration blueprints.
• Connect agent transcripts, failure modes, and evaluation verdicts to harness changes and model adaptation datasets.
• Integrate evaluation suites into CI as merge gates and instrument cold-start times, agent PR merge rates, and review-time economics.
• Maintain human approval gates for specifications, merges, and releases, with regulated code paths remaining human-owned.

Requirements

• Experience owning a development environment, build system, or paved-path workflow used by a multi-team engineering organization.
• Strong Python skills and fluency with containers, Kubernetes, and CI/CD systems at scale.
• Direct experience deploying or operating AI coding agents such as Claude Code, Cursor, Copilot, or comparable in-house systems.
• Knowledge of agentic-systems research, including harness design, reinforcement learning from execution feedback, and evaluation methods.
• Ability to demonstrate measurable developer-experience improvements delivered in production.
• Daily hands-on use of AI coding assistants in development work.

Nice to have

• Experience with hermetic build systems such as Bazel, Buck, or Nix, or monorepo tooling at scale.
• Go or Rust expertise and Git-at-scale experience.
• Experience building unattended agent pipelines with failure caps and human escalation.
• Experience converting agent execution traces into training or evaluation datasets or building reinforcement-learning pipelines.
• Financial-services engineering experience or public writing about developer experience and agent harnesses.

Culture & Benefits

• Small, senior engineering organization where engineers write specifications, harnesses, evaluations, and guardrails while AI agents execute implementation loops.
• Review is prioritized over typing, with no separate QA function or scrum masters.
• Paid time off based on employee grade, including 12–25 vacation days, company holidays, personal days, and sick leave.
• Medical, dental, and vision coverage, life and disability insurance, and employee assistance programs.
• Retirement savings plans such as 401(k) in the United States and additional benefits subject to local policy and eligibility.