Want to know if this job is worth applying to?
San Francisco, California, United States
Onsite
Available
Mission: Give every human superintelligence, and make superintelligence more human.
Intelligence is the product lab company behind DesignArena - 5.2M+ users in 8 months, leaderboard referenced by Andrew Ng, Elon Musk, Demis Hassabis, and more. The team is incredibly talent-dense (11 from Harvard + Berkeley), backed by Tier 1 VC Index Ventures, YC, SV Angel, Lenny Rachitsky, Paul Graham, Figma Ventures, and one of the fastest-growing seed-stage startups in SF.
Behind closed doors, we see what the models can do six months before the world does. We are trusted by the best frontier model providers like OpenAI to rigorously evaluate the capabilities of state-of-the-art multimodal models across design, web dev, game dev, image, video, audio, slide generation, and more, through the large-scale platforms that we’ve built.
Design Arena, our flagship product, is the most referenced benchmark for AI-generated visuals, and is powered by over 5.3M+ authentic users across 192 countries. Prediction Arena was the first time models traded autonomously with real cash on real-time, real-world events. Social Arena tested whether AI models can effectively grow and engage audiences on X by having them operate as independent social media agents.
You'll invent new ways to detect and scale the data that makes frontier AI models smarter. As frontier models continue to improve, model capability is increasingly driven by the quality of training data. You'll develop scalable systems that generate, curate, and improve high-quality supervision across coding, multimodal, and agentic tasks.
In this role, you also have an opportunity to be forward-deployed should it interest you, and work directly with the researchers at frontier labs to creatively scale new model capability strategies for improvement.
Train and improve preference, reward, and ranking models from millions of human interactions
Develop infrastructure for large-scale experimentation, model training, and specialize in online evaluation techniques
Design systems that transform human preference data into reliable signals for downstream model evaluation and training
Data-pilled. Understand that data quality is the real bottleneck to superintelligence.
Strong STEM background. You studied Computer Science, Machine Learning, Data Science, Statistics, Math, Engineering, Physics, or a related field.
Experience building ML systems. You've trained preference models, built data pipelines, or worked on ML infrastructure that runs in production.
Interested in how AI learns from human feedback. You're excited by solving problems at the intersection of human-model interaction.
Location: San Francisco, Levi’s Plaza. We sponsor visas and handle relocation.
Work Schedule: Sunday-Friday. Saturdays are yours!
Compensation: Competitive salary + meaningful equity. You'd be joining at the stage when ownership matters most.
