AI Harness Engineers
About The Position
A new AI-native product organization within Capgemini Financial Services is seeking an AI Harness Engineer. This organization builds software products for insurance claims, payment operations, and health operations, sold to banks, insurers, and health plans. The engineering model is agentic, where AI agents perform most of the implementation, guided by human engineers who author specifications, tooling, evaluation suites, and guardrails. Humans retain ownership of all consequential decisions, with some decisions being human-only by design in regulated domains. The AI Harness Engineer will own the end-to-end development loop, which is critical for the organization's velocity. This includes development environments, builds, CI, and the harnesses that AI agents run in. The goal is to minimize any waiting time for engineers or agents to ensure product shipment. The harness is designed as a self-improving system, utilizing failures, transcripts, and evaluation verdicts to enhance future versions of the harness and the AI models within it.
Requirements
- Prior ownership of a development environment, build system, or paved-path workflow used by a multi-team engineering organization.
- Strong Python skills.
- Container and Kubernetes fluency.
- Comfort operating CI/CD systems at scale.
- Direct experience deploying or operating AI coding agents (e.g., Claude Code, Cursor, Copilot, or in-house), beyond personal use.
- Close following of frontier agentic-systems research (harness design, reinforcement learning from execution feedback, evaluation methods) with the ability to put it into production within the quarter it lands.
- A measurement habit: ability to show numbers for a developer-experience improvement that has been shipped.
- Daily, hands-on use of AI coding assistants as part of your own development workflow.
Nice To Haves
- Hermetic build systems (Bazel, Buck, Nix, or similar) or monorepo tooling at scale.
- Go or Rust experience.
- Git-at-scale experience.
- Experience building one-shot or unattended agent pipelines with hard failure caps and human escalation.
- Experience turning agent execution traces into training or evaluation datasets, or building reinforcement learning pipelines from execution feedback.
- Publicly written about developer experience or agent harnesses.
- Financial services engineering exposure (banks, insurers, or payment providers).
Responsibilities
- Own development environments end to end, ensuring they are fast, isolated, reproducible, and suitable for both human engineers and agent fleets, including sandboxes, ephemeral environments, and warm starts.
- Own deterministic CI and the pre-push validation surface to catch failures early.
- Own the agent harness, including tooling, permissions, retry limits, and orchestration blueprints for coding agents, ensuring permanent improvement with each agent failure.
- Own the recursive improvement loop, ensuring agent transcripts, failure modes, and evaluation verdicts automatically feed back into harness changes and model adaptation datasets.
- Own evaluation-gated merges, integrating evaluation suites as a first-class merge gate.
- Own measurement of key metrics such as cold-start times, agent PR merge rates, and review-time economics, and drive improvements based on this data.
Benefits
- Paid time off based on employee grade (A-F), depending on grade: Vacation: 12-25 days.
- Company paid holidays.
- Personal Days.
- Sick Leave.
- Medical, dental, and vision coverage.
- Retirement savings plans (e.g., 401(k) in the U.S., RRSP in Canada).
- Life and disability insurance.
- Employee assistance programs.
- Other benefits as provided by local policy and eligibility.
Stand Out From the Crowd
Upload your resume and get instant feedback on how well it matches this job.
What This Job Offers
Job Type
Full-time
Career Level
Senior
Education Level
No Education Listed
