SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Intelligence is a product lab company building frontier AI evaluation and training infrastructure. The company operates DesignArena (5.2M+ users, referenced by Andrew Ng and Elon Musk), Prediction Arena, and Social Arena—platforms that rigorously evaluate state-of-the-art multimodal models and generate high-quality training data for frontier model providers like OpenAI.
As a Member of Technical Staff in Post-Training Research & Data, you will invent scalable systems to detect and curate data that improves frontier AI models. The core insight: model capability is increasingly bottlenecked by training data quality, not compute. You'll develop infrastructure to generate, curate, and improve high-quality supervision across coding, multimodal, and agentic tasks.
Key responsibilities:
- Train and improve preference, reward, and ranking models from millions of human interactions
- Develop infrastructure for large-scale experimentation, model training, and online evaluation techniques
- Design systems that transform human preference data into reliable signals for downstream model evaluation and training
- Optionally, forward-deploy to work directly with researchers at frontier labs on new capability scaling strategies
You should have a strong STEM background (CS, ML, Data Science, Statistics, Math, Engineering, Physics), with hands-on experience building ML systems in production—preference models, data pipelines, or ML infrastructure. Deep interest in how AI learns from human feedback is essential.
The company is backed by Tier 1 VCs (Index Ventures, YC, SV Angel) and is one of the fastest-growing seed-stage startups in SF. The team is talent-dense (11 from Harvard + Berkeley). Work schedule is Sunday–Friday with Saturdays off. The company sponsors visas and handles relocation.
