About the opportunity

• Our hiring partner is a YC-backed fintech company building banking and payment automation for the construction industry.
• The company is looking for an experienced engineer to own the internal AI systems already used by its engineering, fraud, customer support, and operations teams.
• This is a hands-on production engineering role, not developer relations, training, or a prompt-only position.

What you will own

• Maintain, harden, and extend an internal AI harness, including evaluations, tool integrations, agent skills, sandboxes, and reliability controls.
• Own AI-powered back-office systems used in fraud and customer-support workflows, working directly with operations when case handling needs improvement.
• Standardize engineering environments across repositories, virtual environments, VMs, secrets, language toolchains, agent runtimes, and development tools.
• Keep shared agent skills and repo instructions accurate, versioned, reviewed, and retired when they become stale.
• Prepare repositories for both engineers and coding agents through reproducible environments, task runners, test entry points, linting, fixtures, and clear contracts.
• Build lightweight jobs that surface daily signals on delivery, review quality, test health, agent usage, and recurring engineering bottlenecks.
• Introduce practical, low-friction processes for pull requests, agent-assisted reviews, skill promotion, and quality control.
• Work closely with engineers and operations users, investigate failures firsthand, and improve the harness, tooling, or back-office software.

What we are looking for

• 6+ years of professional software engineering experience shipping production systems.
• Evidence of hands-on work in this area, such as an agent harness, evaluation framework, reusable agent-skill system, or meaningful open-source contribution.
• Experience setting up reliable development environments and repositories for other engineers, not only for personal use.
• Strong practical engineering ability in Python, TypeScript, or comparable languages, plus CI, automation, scrapers, tool servers, sandboxes, and operational interfaces.
• A production mindset for AI harnesses, agent skills, evaluations, and back-office workflows: versioned, tested, reviewed, monitored, and maintained.
• Comfort operating independently in a small team without a large platform or developer-productivity organization.
• Strong written and spoken English and the ability to collaborate directly with engineering and operations stakeholders.

Bonus points

• Experience rolling out shared Cursor, Claude Code, Codex, or similar agent-assisted engineering workflows across a team.
• Published evaluations, software-engineering benchmark loops, or sandbox and VM runners.
• Contributions to coding-agent or agent-framework ecosystems.
• Experience bringing AI into fraud, customer-support, or other operational workflows.
• Fintech, payments, banking, or another regulated-industry background.

Location and work authorization

• Remote within the United States or Canada, with an in-person option in Austin, Texas.
• Canadian citizens who qualify for TN status may be considered for the U.S. location.
• No other visa sponsorship is available; candidates must meet the applicable work-authorization requirements.

Compensation and benefits

• Canada remote: CAD 150,000–250,000.
• US remote: USD 125,000–175,000.
• Equity: 0.25%–0.50%, with flexibility between cash and equity for the right person.
• Medical, dental, and vision coverage, 401(k), reasonable paid time off, and equipment.

Application and interview process

• Please include a resume, GitHub or LinkedIn profile, and a direct link to a relevant harness, evaluation project, agent-skill system, or open-source contribution.
• Include a short note describing what you have shipped, how you made engineers or operations teams more effective, and what you would inspect first in an existing AI harness and back-office environment.
• The process includes a brief founder screen, a working session focused on a harness, back-office, or environment problem, conversations with engineers, and references.