Documentation Home

Welcome to the Dynamo AI Documentation Page!

Dynamo provides a comprehensive solution for addressing security and compliance requirements for AI models and systems.

Dynamo's product line align to the following key steps of the enterprise AI productionization workflow:

Streamlined Red-Teaming for AI Risks: DynamoEval

DynamoEval provides automated stress testing of AI systems and autogenerates documentation needed for regulatory and risk audits.

• Hallucination Testing
• Security Vulnerability Testing
• Privacy and Data Leakage Testing
• Model Toxicity and Bias Testing

AI Risk Remediations: DynamoEnhance

DynamoEnhance remediates identified risks and enhances models to bolster data privacy, security, and overall robustness of your AI systems.

• RAG Optimizations
• PII Masking
• Model Safety Alignment (i.e. RLHF)
• Differentially Private Fine-Tuning
• Federated Learning

AI Guardrails and Observability: DynamoGuard

DynamoGuard enables enterprises to deploy customizable AI guardrails and offers a full observability platform to audit LLM usage.

• Custom Content Guardrails (i.e. Avoid Legal Advice, Avoid Financial Advice, Block Compensation Reporting, etc.)
• PII Entity Sanitization and Masking
• Toxicity Guardrailing
• Hallucination Detection

How do I get started?

• Test an a RAG pipeline for Hallucinations
• Test your AI system for PII leakage from malicious prompts
• Create a custom Guardrail Policy in natural language