Work

Featured Work

Supporting Evidence

Agent Platform Data Feedback Loop

Supporting evidence for a production Agent platform: connecting traces, badcases, evaluation samples, and preference data.

  • Preference Data
  • DPO
  • Evaluation
Read more
Framework

Production-grade Agent Platform Architecture

A focused evidence framework for Agent runtime, tool systems, permission audit, replay evaluation, and enterprise integration.

  • Agent Runtime
  • Tool-use
  • LLMOps
  • Evaluation
Read more