
Osmosis
San Francisco, CA
Reinforcement learning platform for AI agents
About
Osmosis provides a reinforcement fine-tuning platform that enables companies to train task-specific AI models outperforming foundation models at lower cost. It supports developers building AI agents by handling post-training workflows including feature engineering, reward functions, and real-time learning from experience. This addresses challenges like edge cases and multi-step failures in AI agents.
Tech stack
Python, FastAPI, Golang, React, TypeScript, Next.js, AWS Fargate, Docker, Kubernetes, AWS SageMaker, PyTorch, DynamoDB, S3
Perks & benefits
Competitive salary
More AI companies
AI companiesOpen jobs
1Evaluates AI-generated medical content for accuracy, completeness, relevance, and educational usefulness. Requires an MBBS, MBChB, or MD and at least two years of medical education experience while currently teaching medical students in the UK, New Zealand, or Australia.