Member of Technical Staff, Cortex Lab (applied AI research)

A hands-on applied research and engineering position in Cortex Lab, focused on AI agents and LLM systems that run inside real enterprise workflows. The staff member designs evaluations, benchmarks, and datasets, digs into why agents fail in production, and runs quick experiments to make them more reliable. The role works with researchers, engineers, domain experts, and enterprise customers, and builds Python evaluation pipelines and research tooling.

What you would do

• Design evaluation frameworks, benchmarks, and datasets for LLM agents and complex AI workflows
• Find, analyze, and rank model, agent, and workflow failures seen in production
• Run fast experiments to test hypotheses and improve agent reliability
• Work with enterprise customers and researchers to evaluate, deploy, and tune AI systems
• Build Python evaluation pipelines, experiment tools, and lightweight research infrastructure

What the listing asks for

• Strong background in applied AI research, research engineering, machine learning, or LLM-based systems
• Hands-on experience designing evaluations for AI agents, including tool use and multi-step workflows
• Strong Python skills and experience building evaluation pipelines, experimental tooling, or ML infrastructure
• Ability to trace agent failures to root causes and turn findings into measurable improvements
• Preferred: production agent evaluation, reinforcement learning or post-training familiarity, or frontier lab experience

Suits applied AI researchers and research engineers who like diagnosing why agents break in real business settings and working directly with enterprise customers.

Listed pay: $200,000 to $250,000 a year. Where: Remote. Time: Full-time. Full-time employee role on micro1's own team.

Applications go through micro1's own site. AI Trainer Guide is an independent guide to AI training work, not micro1 or its recruiter. The apply link on our page is a referral link: micro1 may pay us if it hires you, at no cost to you.