Agents, support workflows, onboarding intelligence, human-facing tools, and execution systems for teams that need to move from "AI matters" to "this is live."
I work where human judgment meets practical AI infrastructure. The through-line is simple: replace vague transformation talk with tools, rituals, and operating models people can actually use.
| Track | What It Means | Examples |
|---|---|---|
| Human OS | Tools for the person in the loop: reflection, resilience, writing, health, privacy, and judgment. | Vessel, Gram Reaper, Emotions Soup |
| AI Stack | Agents and interfaces that turn scattered context into grounded action. | Flux, Ground Truth, agent tooling |
| Execution | Operating models for customer-facing teams, FTE planning, and the AI transition. | The Support Machine, barnirhired, GiFOC, workshops |
| Project | What It Is | Stack |
|---|---|---|
| the-support-machine | A free book and executable AI-agent skill for deploying AI in customer support without breaking trust. | Python, AI support, operating model |
| flux-hackathon-2026 | Live onboarding intelligence for GitLab teams, built with Google ADK, Gemini, GitLab MCP, Notion, and Cloud Run. | Python, Gemini, GitLab MCP |
| Vessel | A private therapeutic writing environment inspired by EMDR, with local encryption and adaptive atmosphere. | TypeScript, Tauri |
| Conflict 2027 | A 48-turn political simulation game where peace is winnable and genuinely hard, with deterministic game balance and optional Claude-authored briefings, crises, and negotiation counterparts. | TypeScript, Next.js, Claude, simulation |
| gram-reaper | An open-source protein tracker with natural-language AI parsing, quick-add cards, streaks, and PWA install. | Vue, SQLite, Claude |
| ground-truth | One topic, five editorial perspectives, no feed and no algorithmic addiction loop. | Next.js, xAI |
| logo-v2 | A remake of the generational Logo turtle educational system from the beginning of personal computing. | JavaScript, Logo, education |
| ada-revival | An in-browser Ada/SPARK playground and MCP kit for proving code correct and helping AI assistants write safer systems software. | TypeScript, Ada, SPARK, MCP |
| ada-quest | Interactive Ada and SPARK missions for learning systems safety through small bugs. | JavaScript, Ada, SPARK |
- AI support without broken trust
- Grounded agents that read live systems, cite sources, ask for confirmation, and perform useful work
- Proof-oriented systems: memory safety, SPARK verification, and AI assistants that can work in under-supported technical domains
- Human agency tools that protect privacy, judgment, emotion, and ownership while AI gets stronger
- barnirhired.com - job-fit analysis for people applying to technical, customer-facing roles like Solutions Architect, Customer Success, Sales Engineering, Forward Deployment, TAM, and Support Engineering
- A growing portfolio of shipped products, experiments, books, and live tools across AI, support, education, and human systems
- 4 books across AI, leadership, and human systems
- Languages: TypeScript, Python, JavaScript, HTML/CSS, Go, Rust, Ada/SPARK
- Frontend/app: React, Next.js, Vue, Astro, Vite, Tailwind, Tauri, Tone.js
- Backend/AI/infra: FastAPI, Express, Node.js, SQLite, Zod, Vitest, Docker, Cloud Run, Vercel, GitHub Pages, MCP/GitLab MCP, Google ADK, OpenAI, Gemini, Claude/Anthropic, xAI