AI Safety Benchmark Engineer

White Circle is seeking a research engineer to build and maintain our internal benchmark suite for single/multi-turn content and agentic safety, collaborating with product on evals for our flagship models. You will also study agent behaviours in the wild and contribute to research projects.

You will write production-grade Python, design scalable benchmarks, and extend evals to new data, drawing on experience creating LLM benchmarks and synthetic data, with strong experimental rigor and

#J-18808-Ljbffr…