Remote Contract: Agentic Search Systems Engineer

Mercor; San Francisco (CA); Remote; GBP 83,000 - 156,000; Part time

Job summary

Mercor is hiring a Software Engineer for Agentic Search Systems to help improve AI-driven search by evaluating AI-generated content and reinforcing model rigor. The role focuses on building and scaling agentic search capabilities, with a strong emphasis on data infrastructure to support efficient retrieval and reliable performance.

The ideal candidate ships production search systems, owns relevance and retrieval for real users, and can quantify improvements across models and systems.

Qualifications

• Experience shipping production search systems.
• Ownership of relevance or retrieval on systems used by real users.
• Ability to quantify impact and improvements in search systems.
• Experience building systems related to agentic search.
• Experience scaling data infrastructure for search systems.

Responsibilities

• Evaluate the accuracy and depth of AI-generated content in search systems to strengthen reasoning and rigor in model outputs.
• Develop and optimize systems related to agentic search for improved performance and reliability.
• Scale data infrastructure to support modern search systems and enhance retrieval capabilities.
• Collaborate with AI research teams to quantify impact and improvements in search systems.
• Work independently and asynchronously to meet deadlines while improving AI model performance.

Skills

Production search systems

Relevance retrieval ownership

Quantify impact in search

Agentic search development

Data infrastructure scaling

Job description

Mercor is hiring a Software Engineer for Agentic Search Systems to help improve AI-driven search by evaluating AI-generated content and reinforcing model rigor. The role focuses on building and scaling agentic search capabilities, with a strong emphasis on data infrastructure to support efficient retrieval and reliable performance.

The ideal candidate ships production search systems, owns relevance and retrieval for real users, and can quantify improvements across models and systems.