Location: San Francisco Bay Area, CA
Remote: Yes (US and international consulting); available for occasional onsite work and travel
Willing to relocate: No
Technologies: LLM agents; Claude Code; Codex; MCP; multi-agent orchestration; evals; workflow automation; legal-AI architecture; privilege and authority controls; system-prompt assurance
Résumé/CV: <a href="https://www.dazzagreenwood.com/about" rel="nofollow">https://www.dazzagreenwood.com/about
LinkedIn: <a href="https://www.linkedin.com/in/dazzagreenwood" rel="nofollow">https://www.linkedin.com/in/dazzagreenwood
Email: dazza@civics.com
SEEKING CONSULTING / CONTRACT WORK — also open to the right full-time role
Lawyer-technologist designing and building agentic AI systems for legal and other high-stakes work. I focus on the difficult distance between a promising demo and a deployment-ready workflow: selecting the right work, mapping privilege, data, and authority boundaries, designing evals and human-decision gates, and creating tool-agnostic operating architectures that can survive legal and security review.
Current obsession: chaining authority- and scope-bounded agents into high-assurance workflows that can run autonomously for six-plus hours, show their receipts, and stop at explicit human (sometimes licensed professional) gates between stages—then produce genuinely valuable, substantial, compliant, and defensible deliverables.
A useful first engagement is a bounded two-week sprint: rank workflows by value, readiness, and fit; sketch an initial target operating architecture; and produce a prioritized 90-day pilot plan plus one implementation-ready pilot brief.
Recent work:
AISPA system-prompt assurance paper: <a href="https://arxiv.org/abs/2607.28617" rel="nofollow">https://arxiv.org/abs/2607.28617
Authority Boundaries for AI: <a href="https://www.dazzagreenwood.com/p/authority-boundaries-for-ai" rel="nofollow">https://www.dazzagreenwood.com/p/authority-boundaries-for-ai
Interlateral Agents: <a href="https://github.com/dazzaji/interlateral_agents" rel="nofollow">https://github.com/dazzaji/interlateral_agents (I have a private repo version with professional authorization or approval/adoption gates and control kernel verifiable logging)
Earlier work includes founding law.MIT.edu, testifying before Congress on electronic-transactions law, and leading Consumer Reports’ Data Rights Protocol. I am currently a Stanford CodeX affiliate.
Available now for one bounded sprint and booking one or two larger engagements beginning in September. Best fit: in-house legal teams, legal-tech companies, and AI companies building high-stakes professional workflows. Contract through <a href="https://www.civics.com" rel="nofollow">https://www.civics.com.
dazzaji · · focus · HN ↗
SEEKING CONSULTING / CONTRACT WORK — also open to the right full-time role
Lawyer-technologist designing and building agentic AI systems for legal and other high-stakes work. I focus on the difficult distance between a promising demo and a deployment-ready workflow: selecting the right work, mapping privilege, data, and authority boundaries, designing evals and human-decision gates, and creating tool-agnostic operating architectures that can survive legal and security review. Current obsession: chaining authority- and scope-bounded agents into high-assurance workflows that can run autonomously for six-plus hours, show their receipts, and stop at explicit human (sometimes licensed professional) gates between stages—then produce genuinely valuable, substantial, compliant, and defensible deliverables.
A useful first engagement is a bounded two-week sprint: rank workflows by value, readiness, and fit; sketch an initial target operating architecture; and produce a prioritized 90-day pilot plan plus one implementation-ready pilot brief.
Recent work: AISPA system-prompt assurance paper: <a href="https://arxiv.org/abs/2607.28617" rel="nofollow">https://arxiv.org/abs/2607.28617 Authority Boundaries for AI: <a href="https://www.dazzagreenwood.com/p/authority-boundaries-for-ai" rel="nofollow">https://www.dazzagreenwood.com/p/authority-boundaries-for-ai Interlateral Agents: <a href="https://github.com/dazzaji/interlateral_agents" rel="nofollow">https://github.com/dazzaji/interlateral_agents (I have a private repo version with professional authorization or approval/adoption gates and control kernel verifiable logging)
Earlier work includes founding law.MIT.edu, testifying before Congress on electronic-transactions law, and leading Consumer Reports’ Data Rights Protocol. I am currently a Stanford CodeX affiliate.
Available now for one bounded sprint and booking one or two larger engagements beginning in September. Best fit: in-house legal teams, legal-tech companies, and AI companies building high-stakes professional workflows. Contract through <a href="https://www.civics.com" rel="nofollow">https://www.civics.com.