Open source projects I'm currently working on. See also my resume and GitHub.

  • HumaneBench

    Evaluation framework for measuring human-friendly behavior in AI assistants, built on AISI Inspect. Tests AI responses across humane technology principles including loneliness, mental health, transparency, and privacy. Supports multiple LLM providers with extensible scenario datasets.

    Python

  • Scan MCP

    MCP (Model Context Protocol) server for scanner automation and document digitization. Privacy-first architecture with smart device discovery, JSON Schema validation, and support for both local stdio and network HTTP transports.

    TypeScript, Node.js