Software Engineer (Codebase Deep Reasoning & Evaluation)

About The Role Mercor is seeking software engineers to support one of the world’s leading AI labs in advancing code understanding and reasoning capabilities for next-generation machine learning models. In this role, you’ll engage in real-world engineering work: analyzing large, production-grade repositories to create and evaluate technically challenging coding questions. You’ll systematically explore multiple modules, connect related functions across files, and assess how advanced AI systems reason about architecture, data flow, and performance. Your ability to reason from evidence: citing specific files, functions, and line numbers will directly influence how these AI models learn to think like expert engineers. ⸻ You’re a Great Fit If You Have 4+ years of elite software engineering experience at top-tier startups, quantitative trading firms, hedge funds, or similar high-performance environments. Have experience using coding agents or LLMs as part of your engineering workflow (e.g., Copilot, Claude, GPT-4, or Replit Agents). Hold a Computer Science degree from a leading university or equivalent practical expertise. Are fluent in Python and JavaScript/TypeScript, and can comfortably read Java, Go, or other modern languages (Rust, C++, C#). Demonstrate systematic exploration, you examine multiple files and dependencies before forming conclusions. Practice evidence-based reasoning, grounding your answers in specific code references rather than assumptions. Excel at cross-file synthesis, connecting distributed logic to explain how systems work end-to-end. Show strong architectural understanding, identifying patterns, abstractions, and design choices in complex codebases. Display intellectual honesty: you acknowledge uncertainty when information is incomplete or ambiguous. Write clear, structured technical documentation, and communicate insights precisely and persuasively. ⸻ Example Projects & Domains You may work across diverse systems, including: Web APIs and backend services CLI tools and data processing pipelines Frontend applications and DevOps tooling Security, observability, and performance-critical architectures Each task will challenge your ability to connect architecture, dependencies, and logic across real-world repositories. ⸻ Engagement Details This project will be a high-impact 24-hour sprint launching in the next 1–2 weeks. Compensation: Task-based pay (top performers previously earned $1,000+ during the sprint) Classification: Hourly contractor through Mercor Payment: Weekly payouts via Stripe Connect ⸻ About Mercor Mercor connects elite creative and technical talent with leading AI research labs, headquartered in San Francisco, CA. Our distinguished investors include Benchmark, General Catalyst, Peter Thiel, Adam D’Angelo, Larry Summers, and Jack Dorsey. Apply today and redefine digital creativity alongside the teams building the future of intelligent software.