Summary
Surge AI is building a platform that supports the development, evaluation, and improvement of advanced AI models through human expertise and scalable oversight. The Data Scientist will partner with AI labs to design data collection strategies, analyze experimental results, and improve model performance, safety, and alignment.
Responsibilities
- Design data collection workflows optimized for particular model needs
- Implement quality controls to reduce biases and ensure model safety
- Run experiments to make leading models more insightful, compelling, safe, and human aligned
Skills
- Passion for Data Analysis - Analyzing data and iterating on collection methods to improve model performance
- Curiosity About LLMs - You've dug into how large language models work — not just as a user, but at a conceptual level. You can speak confidently about fine-tuning, RLHF, evaluation, and model behavior, even if you haven't built one from scratch
- Love of Customer Collaboration - You're energized by working directly with AI researchers and technical customers. You can translate complex model behavior into clear insights, ask good questions, and earn trust in a room full of experts
- Python Proficiency - You're comfortable writing Python to analyze data, build pipelines, and prototype workflows. You don't need to be a software engineer, but code is one of your primary tools
Qualifications
Must Haves
- Passion for Data Analysis - Analyzing data and iterating on collection methods to improve model performance
- Curiosity About LLMs - You've dug into how large language models work — not just as a user, but at a conceptual level. You can speak confidently about fine-tuning, RLHF, evaluation, and model behavior, even if you haven't built one from scratch
- Love of Customer Collaboration - You're energized by working directly with AI researchers and technical customers. You can translate complex model behavior into clear insights, ask good questions, and earn trust in a room full of experts
- Python Proficiency - You're comfortable writing Python to analyze data, build pipelines, and prototype workflows. You don't need to be a software engineer, but code is one of your primary tools
Benefits