AI-Native Systems Engineer for Drug Discovery

Mirror Physics Corporation; New York (NY); On-site; USD 150,000 - 230,000; Full time

Benefits offered by this job

Equity

Health benefits

Fitness budget

Unlimited PTO

Job summary

Mirror is an NYC-based startup building the AI stack for modern drug discovery. We develop agents that give scientists the leverage to explore, test, and advance new medicines faster and at lower cost.

Our platform bridges frontier model capabilities with practical preclinical challenges to accelerate scientists’ daily work with transparency and data security. As a lead engineer, you’ll help us build a high-quality platform for experts and the surrounding systems that enable scale without

Responsibilities

• Design, architect, and ship production systems across the stack, including distributed compute, databases and data pipelines, backend services, user-facing applications, and observability infrastructure.
• Build and improve multi-agent systems, agent harnesses, tool integrations, and inference infrastructure.
• Develop internal and external benchmarks and evaluations that measure scientific quality, reliability, and usefulness.
• Measure and improve agent quality, cost, latency, and robustness.
• Build repeatable pipelines for integrating and verifying new tools, with traceability for scientific work.
• Create feedback loops that turn evaluation results into systematic agent improvements and keep benchmarks current.
• Ship new harness, interface, and enterprise capabilities, and automate high-leverage internal workflows as priorities evolve.

Job description

Mirror is an NYC-based startup building the AI stack for modern drug discovery. We develop agents that give scientists the leverage to explore, test, and advance new medicines faster and at lower cost.

Our platform bridges frontier model capabilities with practical preclinical challenges to accelerate scientists’ daily work with transparency and data security. As a lead engineer, you’ll help us build a high-quality platform for experts and the surrounding systems that enable scale without