Staff AI Scientist
Core
Lead applied research and development for models and datasets powering Fiddler's Trust Service, guardrail classifiers, and evaluators to ensure LLM and agentic applications are safe, accurate, and compliant in production.
Role type
Staff AI Scientist (Applied Research & Production)
Builds
Production classifiers for safety/security/quality, synthetic/adversarial dataset pipelines, and LLM-powered diagnostic analysis layers.
Domain
Generative AI, AI Observability, Enterprise AI Governance
Deliverable
production ML models
Required skills
Applied AI research to production, LLM/Agentic evaluation and guardrailing, classifier model training (BERT-family, LLM-as-classifier), dataset engineering (sourcing, synthetic generation, adversarial augmentation), Python, PyTorch, Hugging Face, real-time inference optimization, reinforcement learning and preference-based methods
Preferred skills
Publications, talks, or open-source contributions in AI
Technologies
PyTorch, Hugging Face, BERT-family, ModernBERT
Responsibilities
Design and ship production classifiers under latency/cost constraints; lead synthetic and adversarial dataset pipeline development; drive technical direction of generative insights; contribute to evaluation infrastructure for model quality and drift; mentor AI Scientists; partner with engineering and product teams to translate customer needs into research problems.
Seniority
Staff, hands-on IC with mentorship and technical direction