Skip to content
View gianlucamazza's full-sized avatar
:electron:
writing prompts
:electron:
writing prompts

Highlights

  • Pro

Block or report gianlucamazza

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
gianlucamazza/README.md

Gianluca Mazza

Technologist and entrepreneur building intelligent agent systems — I've been writing software since before it was cool.

I design and lead platforms that combine LLM agents, automation frameworks and verifiable infrastructure, and I ship them the boring way: ADR before code, falsifiable claims, everything self-hosted on Linux. Currently co-founder of an AI research and engineering venture; previously head of AI and CTO roles across a few companies.

Stack

  • AI & agents — agent orchestration, MCP servers, local LLM inference (llama.cpp / GGUF, ONNX Runtime GenAI), RAG pipelines
  • Languages — Rust, Python, TypeScript, modern C++
  • Bitcoin & crypto — Bitcoin Core, Solana / Jupiter, zero-knowledge proofs
  • Systems & infra — Linux, systemd, Docker, Kubernetes, GitHub Actions

Selected work

Agents & MCP

  • orka — modular agent orchestration platform in Rust: priority queues, skill system, MCP/A2A, sandboxed execution
  • reasoning-kernel — make prompt injection structurally unable to cause an unauthorized effect: a CaMeL-style capability and taint-tracking reference implementation
  • mcp_python_toolbox · mcp-duckduckgo · mcp-ascii-charts — MCP servers for Python development, web search and terminal charts
  • agentroom — end-to-end encrypted agent-to-agent chat (Double Ratchet) over msg2agent, the blind Go relay underneath

Systems & low-level

  • xllama — local LLM chat and Stable Diffusion on an Xbox Series S|X: llama.cpp (GGUF) and ONNX Runtime GenAI, CPU and DirectML
  • harbor-kernel — verified bare-metal Rust kernel for the Raspberry Pi 4 (AArch64): cooperative tasks, EL0 agents, IPC, W^X MMU
  • uwp-crossbuild · openappx — compile, package and sign Windows/UWP apps entirely from Linux; the toolchain that makes the two Xbox projects possible

Bitcoin & finance

  • solana-mmaker — market-making bot automating strategies through the Jupiter swap protocol
  • xbox_bitcoind — pruned Bitcoin Core full node running on Xbox Series S|X developer mode
  • openfatture — CLI-first open-source electronic invoicing (FatturaPA) for Italian freelancers

Elsewhere

gianlucamazza.it · LinkedIn · X

Pinned Loading

  1. emotional-memory emotional-memory Public

    Emotional memory for LLMs based on Affective Field Theory (AFT)

    Python 4

  2. agentroom agentroom Public

    Agent-to-agent encrypted chat over a self-hosted relay. E2E encrypted, blind server, Double Ratchet.

    TypeScript

  3. xllama xllama Public

    Local LLM chat and Stable Diffusion on Xbox Series S|X (UWP Dev Mode) — llama.cpp (GGUF) + ONNX Runtime GenAI, CPU and DirectML.

    C++ 9 1

  4. solana-mmaker solana-mmaker Public

    Automated market making on Solana: runs configurable spread strategies through the Jupiter swap protocol.

    TypeScript 96 33

  5. openfatture openfatture Public

    Open-source electronic invoicing (FatturaPA) for Italian freelancers, CLI-first.

    Python 2 2

  6. mklang mklang Public

    A declarative language for LLM-driven state machines - the .mkl document is the program, an LLM is the runtime

    Python 1