Digging into agent orchestration internals: hand-rolled state machines vs LangGraph, RAGAS-driven eval loops, and token-efficient tool design for LLM harnesses like Claude Code.
Currently hacking on
repi, an autonomous log investigation agent (hybrid retrieval, state-machine loop, RAGAS-evaluated across 5 models), and smrit, token-efficient MCP tooling for LLM coding harnesses.
Available for
Chats on production RAG, agent orchestration, retrieval evaluation, or Go backend infra. Also open to backend / AI infra roles in Bangalore. Say hi if any of that overlaps.
0 posts published
0 comments written
23 tags followed
loading...
We're a place where coders share, stay up-to-date and grow their careers.