Kev: Tiny Jev-like family of decision models built on top of Qwen3.5

Kev introduces a small, self-hostable family of decision models built on Qwen3.5 that can classify, score, and route inputs with probability outputs, making them well-suited for customer support triage, workflow automation, and other structured decision tasks. For CIOs, the strategic value is lower-cost, privacy-preserving AI that can run on local infrastructure or even Apple Silicon, reducing dependence on external APIs while improving control over data, latency, and integration. IT organizations should view this as a practical option for embedding AI into operational decisioning, but one that still requires careful benchmarking for accuracy, calibration, and governance before broad adoption.