
Pareto AI
San Francisco, CA
Verification layer for frontier AI reinforcement learning
About
Pareto AI builds infrastructure to turn expert human judgment into scalable reward signals for training frontier AI models. They provide expert-labeled data, evaluation pipelines, and verification services to AI labs like Anthropic and Google DeepMind. This enables more capable and safer AI models by leveraging real-world expertise.
Tech stack
PostgreSQL, ClickHouse, Python, Django, React, AWS, TypeScript, Agentic Workflows, GCP, Terraform, RAG, SQL, LangChain, Celery, Redis, Fargate, ECS
More AI companies
AI companiesOpen jobs
3Owns reinforcement learning environments end to end, from scoping research goals and building reproducible containerized tools and graders through deployment and production operations. Requires 7+ years of production engineering experience, strong Python or TypeScript skills, cloud and container expertise, and close stakeholder collaboration.
Maintains the physical office environment, kitchen, supplies, deliveries, and occasional events for a busy San Francisco team. The role requires reliability, initiative, attention to detail, independent work, and a commitment to at least 10 hours per week across three days.
Manages strategic AI lab accounts as trusted advisor, owning relationships, solution design for ML workflows like RLHF and data labeling, renewals, expansions, and health metrics for long-term success.