My focus lies in real-world RAG tooling, token telemetry, and observability for vector-based systems. Think of it as developer tools for semantic relevance — but with deep diagnostics, not just dashboards.

Always exploring scalable infra patterns, agent-core workflows, and fine-grained control over ranking signals.

Let’s connect if you’re into LLM-native retrieval, vector observability, or want to discuss what makes a “relevant” result actually relevant.