About the job
Join our team as a Data Scientist and play a key role in managing and optimizing Large Language Models (LLMs) for AI safety.
About Us
White Circle is a pioneering AI Safety company dedicated to enhancing the safety, reliability, and efficiency of AI systems. Our platform is built on policies—clear natural language rules that dictate the appropriate conduct of AI models. We automate the testing, enforcement, and continuous enhancement of these policies at scale.
We have secured $11 million in funding from leading investors, including founders and senior leaders from OpenAI, Anthropic, HuggingFace, Mistral, DeepMind, Datadog, Sentry, and others.
Our operations handle over 100 million API calls each month.
We specialize in fine-tuning and training our own LLMs, enabling them to operate faster and at a lower cost than any available model.
As a compact and dedicated team, we offer an environment where you can tackle complex challenges, expedite your work into production, and substantially influence the development of AI safety.
Your Role:
Transform vast amounts of unstructured text into a structured and navigable format (including topics, clusters, segments, trends, and anomalies), moving from “unknown unknowns” to stable definitions for tracking.
Develop scalable representation pipelines: including sampling strategies, preprocessing/normalization, embeddings at scale, indexing, and retrieval, making the corpus searchable and analyzable.
Utilize LLMs in practical applications: implementing labeling/classification, weak supervision, data enrichment, summarization, and automated diagnostics of incoming volumes, with cost and quality controls in place.
Provide actionable insights that inform decisions: convert your findings into product and operational strategies (identifying available data, gaps, quality issues, and prioritization).
Create self-service analytics resources: develop datasets, data models, and user-friendly tools/dashboards to enable team members to explore and address inquiries independently.
Collaborate closely with engineering and research teams: ensure alignment of pipelines with production constraints (including latency, cost, and privacy) and integrate outputs into existing workflows.
You Will Thrive If You:
Possess strong skills in Python and SQL, with an engineering mindset capable of building reliable pipelines rather than just notebooks.
Have practical experience in applied NLP/ML on real-world text: including embeddings, clustering, and model training.

