Senior backend engineer building LLM systems that behave in production — Go for the systems, Python for the agents.
I work at the boundary between solid backend infrastructure and language models that have to be reliable: structured output, retrieval, multi-agent orchestration, and the deterministic fallbacks, cost caps, and validation guardrails that keep them honest. Eight years across Go, Node/TypeScript, and cloud platforms — from backend IC work to leading engineering teams.
CloudOracle — a multi-cloud FinOps toolkit in Go + Python. It audits live AWS/GCP/Azure spend, prices Terraform PRs before they merge (as a GitHub Action), and answers cost questions through a guardrailed multi-agent LLM system: a hand-rolled LangGraph supervisor, RAG over pgvector, semantic answer validation, and deterministic no-LLM fallbacks.
Go · Python · TypeScript / Node · React / React Native · PostgreSQL · pgvector · Redis · AWS · GCP · Docker · LangGraph · FastAPI
- Building agentic LLM systems with real reliability engineering — evals, cost caps, deterministic fallbacks
- Working toward CKAD
- LinkedIn — /in/cro22
- X — @wolfdeerdev




