Skip to content
← All projects

Observability & evals projects

Tracing and evaluation of LLM and agent systems. Built by Chiranjeev Thomas, senior full-stack engineer, agentic ai.

In progress
  • AI engineering

Agent observability & evals for InvoiceChaser

Per-step tracing of every agent run, plus a golden-set eval that checks tone and tool choice at each escalation level before a prompt change ships.

  • Tracing
  • Evals
  • Python