Essays, playbooks, and engineering notes from the DOVIX AI team.
Operations
Most pilots succeed in demo and fail in production. Here's the playbook to bridge the gap.
Engineering
Our opinionated agent architecture: LangGraph, MCP, evals, observability, and human handoff.
Research
Why naive RAG keeps disappointing — and the retrieval architecture that actually works.
Without an eval harness, you don't have a model — you have a vibe.
Strategy
A pragmatic guide to choosing between hosted frontier models and self-hosted open models.
Security
Multi-agent triage, correlation, and response in modern security operations.