Software Engineering Expert

Role Overview

Theta partners with top AI labs to provide high-fidelity training data for the next generation of coding models. We specialize in authentic software engineering workflows: building features in existing codebases, refactoring legacy systems, and fixing critical bugs.

As a Software Engineering Expert at Theta, you'll create the training data that teaches AI models to think like real software engineers. Working with production-scale codebases across multiple languages and frameworks, you'll design complex, multi-file engineering tasks that reflect the messy reality of professional software development. You'll work as an AI engineer—building the evals and datasets that define frontier model capabilities and shape the future of AI-assisted software development.

What You'll Do

• Design realistic software engineering tasks in production-scale codebases (both open-source and private repositories) across multiple languages, frameworks, and domains (backend, frontend, infrastructure, etc.)
• Create novel contributions to open-source repositories, developing new features, refactoring opportunities, and authentic bug fixes that haven't been implemented
• Write comprehensive problem statements with detailed specifications that mirror real-world product requirements and technical constraints
• Build reference implementations demonstrating correct solutions, including handling edge cases and making appropriate architectural trade-offs
• Develop test suites that validate functionality across multiple files and components, ensuring solutions work in the broader system context
• Document human reasoning traces that capture the thought process of an experienced engineer: how to navigate unfamiliar code, debug issues systematically, and make incremental progress on complex tasks
• Review and provide feedback on code and tasks created by other experts to maintain rigorous quality standards, and collaborate on improving our data creation infrastructure

What We're Looking For

• Professional software engineering experience. You've worked in production codebases, shipped features to users, and dealt with real-world engineering complexity (or equivalent demonstrable experience)
• Strong programming fundamentals with proficiency in at least 2-3 languages (Python, JavaScript/TypeScript, Java, C++, Go, or similar)
• Full-stack perspective. You understand how different parts of a system fit together, from databases to APIs to frontend interfaces
• Comfort with ambiguity and messy code. You can navigate unfamiliar codebases, reverse-engineer existing architectures, and work in realistic (imperfect) environments
• Attention to detail. You write clear specifications, anticipate edge cases, and create comprehensive test coverage
• Self-directed and reliable. You can manage your own schedule, meet deadlines, and maintain high standards without constant oversight
• Excellent written communication skills. You can clearly explain technical decisions, document your reasoning, and provide actionable feedback

Preferred Qualifications

• Experience with multiple programming paradigms and frameworks (e.g., React, Next.js, Django, Spring, distributed systems frameworks)
• Significant contributions to open-source projects or experience reviewing pull requests
• Experience working in large, legacy codebases with complex dependencies

Why Theta

• Autonomy and trust: We provide the platform, tools, and objectives—you control how you design tasks. We rely on your judgment to identify compelling problems, where models fall short, and how to push AI capabilities forward.
• Collaborative culture and direct impact: Work closely with our founding team in a high-trust environment with quick feedback loops. Your insights shape our product direction and inform model development at top AI labs.
• Continuous learning: Work across diverse codebases, frameworks, and domains, ranging from React frontends to distributed systems to ML infrastructure. Every project brings new technical challenges.
• Growth opportunities: Start by creating high-quality training data, then take on leadership responsibilities like managing contributors, designing evaluation frameworks, and building internal tooling.