Skip to content

Agents#

This section serves to provide guides and examples for the common usage of agent tasks and agent deployments through the Hopsworks UI and APIs.

  • Agent Tasks: Fire-and-forget agent executions that run as first-class Hopsworks jobs.
  • Agent Deployments: Served conversational agents, the chat contract they answer by, and how to deploy, trace and talk to them.
  • Agent Memory: The four memory tiers a deployment provides, how to configure them, and what a person can see and erase.
  • Feedback: Verdicts on what an agent answered, from reviewers and from end users, and what reads them.
  • Evaluation: Suites of tasks graded by checks, the runs that execute them, and the gate a release reads.
  • Monitoring: Grading the traffic an agent actually served, on a schedule.
  • Failure Analysis: Turning feedback into regression tasks, with a reviewer deciding and a model drafting.