Harvey Feature Trails

This post details the technical process of scaling Harvey's document processing capabilities, focusing on architectural changes, data processing pipelines, and performance optimizations to handle increased load and complexity.

This post introduces the Connector Library, a new system capability designed to centralize context for Harvey's products. It details the architectural design, the process of building the library, and its role in enabling seamless integration and data flow across different Harvey systems. The library aims to provide a standardized way to access and manage contextual information, improving developer efficiency and system interoperability.

This post introduces Harvey's custom Cloud Agent Infrastructure, detailing its architecture, the motivations behind building it in-house, and the technical challenges overcome. It covers the design choices made to ensure scalability, reliability, and efficient operation of agents deployed within cloud environments for various system capabilities.

This post introduces the concept and initial build of an Agentic Security Operations Center (SOC) at Harvey. It details the architecture and technical considerations for integrating AI agents into the SOC workflow to automate threat detection, analysis, and response. Key components discussed include agent orchestration, data ingestion pipelines for security telemetry, and the development of specialized agents for tasks like incident triage and evidence collection. The focus is on building a scalable and intelligent system to enhance security posture.

This post details the technical process of rebuilding Harvey's review algorithm, focusing on architectural changes, data processing, and performance improvements to enhance the system's capabilities.