I build systems that have to survive contact with reality.
Right now that means MUTX: infrastructure for seeing, controlling, and proving what AI agents do in production. Around it, I work on local-first AI, onchain coordination, regulated workflows, market infrastructure, and small terminal programs that are useful for no reason other than joy.
MUTX · GitHub · Docs · X · Telegram · Email
The control plane around the agent. MUTX keeps identity, runs, permissions, budgets, approvals, observability, and audit history in one operating surface.
It ships as more than a dashboard:
- a FastAPI control plane with explicit
/v1/*contracts - an authenticated web operator and guided browser demo
- a signed and notarized macOS app
- a Python CLI, Textual TUI, and first-party SDK
- local and hosted setup paths
- Docker, Terraform, Ansible, Helm, and monitoring assets
- policy enforcement, credential brokering, session budgets, and review gates
Try the product · Read the code · Open the docs · Download for macOS
| Project | What shipped |
|---|---|
| Tablebeam | Private-by-default Q&A for CSV and Google Sheets. Runs against LM Studio, Ollama, or any local OpenAI-compatible model; every answer points back to the source rows. |
| ckitty | A tiny native terminal cat in C11 and ncurses. Procedural poses, deterministic rendering, resize-safe animation, tests, sanitizers, and no runtime baggage. |
| SecurePath | A production-minded Discord research agent for crypto and DeFi: chart vision, contextual conversations, caching, rate control, and operational telemetry. |
| BarterTrade | A full-stack commodity barter platform with real-time offers, identity workflows, AI-assisted matching, and privacy-preserving verification experiments. |
| Terminal Starfield | A dependency-free, interactive 3D starfield for the terminal with warp, trails, density controls, and color modes. |
| MUTX Homebrew Tap | The maintained install path for the MUTX CLI on macOS. |
A very small demo
The repositories are private; the problems are real. This is the part I can describe publicly.
| System | What it does |
|---|---|
| OpenRNA Foundry | A local-only RNA construct review and design-triage workbench. Deterministic sequence scoring, synonymous codon optimization, IVT specification assembly, evidence capture, specialist review agents, and research-PDF workflows—with no telemetry or cloud dependency. |
| Cipher | An autonomous, oracle-aware agent built around programmable onchain behavior, market context, and economic constraints. |
| TCE12 / SecurePath | Cross-border governance and settlement infrastructure: compliance attestations, milestone escrow, royalties, offtake agreements, and proof-of-settlement audit packets. |
| MUTX operations | The private delivery layer behind the public product: runtime operations, deployment infrastructure, authentication, webhooks, API keys, and production control surfaces. |
| Market intelligence systems | Crypto research agents, seed-stage discovery, portfolio and treasury monitoring, vehicle-market data pipelines, and decision-support workflows. |
| AI operator systems | Closed-loop research, qualification, outreach, follow-up, QA, and reporting systems built around measurable commercial outcomes. |
- SecurePath Academy — education and research product
- MUD Escola de Cerâmica — multilingual product and commerce surface
- Cavapendolandia — minimalist art platform
- bilingual hospitality, wellness, and commerce builds
- ambient Discord agents and community intelligence tools
- Loop Courier and Neon Rogue Snake — small game experiments
I came to software through crypto research and institution-facing infrastructure: staking economics, validator governance, custody architecture, MEV, rollups, token design, treasury strategy, and derivatives research.
That work included building SecurePath, institutional staking research at Colossus Digital, quantitative crypto research at Optimum Complexity, co-founding AlphaBlock Network, and contributing tokenomics and governance work to Cerebrum DAO.
It still shapes how I build: explicit incentives, bounded authority, observable state, and systems that can explain what happened after the demo ends.
Languages — Python, TypeScript, Solidity, C, SQL, Bash
Product — FastAPI, React, Next.js, PostgreSQL, Redis, Docker
Infrastructure — Terraform, Ansible, Helm, CI/CD, observability
Human languages — English, Italian, Spanish, Portuguese; basic German
If you are building agent infrastructure, local-first AI, or economically serious onchain systems, get in touch.