Member of Technical Staff - Software Engineering Capabilities

About the Company

Preference Model is automating ML engineering by developing models with advanced software engineering capabilities. The founding team has experience building data infrastructure and datasets for Anthropic's Claude, and they are now partnering with leading AI labs to push the boundaries of frontier models.

Responsibilities

• Hunt for where frontier models break across software and build high-fidelity scenarios to expose failures.
• Own complex roadmap problems end-to-end, including multi-step workflows and large codebases with technical debt.
• Build robust verification systems that prevent models from gaming the evaluation.
• Direct coding agents in daily work and critically evaluate their output.
• Build internal tooling and mentor newer engineers as the team grows.

Requirements

• Deep software engineering experience across multiple domains.
• Expertise in at least one specialty: infrastructure, distributed systems, performance, security, compilers, or databases.
• Proficiency in Python.
• Extensive hands-on experience with coding agents like Claude Code, Cursor, or Codex.
• Strong intuition for model behavior and ability to anticipate shortcuts.
• Proven track record of owning work end-to-end in complex, ambiguous environments.

Preferred Qualifications

• Experience as a senior or staff engineer at a frontier lab or systems-heavy startup.
• Deep specialty expertise in areas where current models struggle.
• Experience as an early engineer at a startup.
• Background in building with coding agents or contributing to agent evaluation work.