Skip to content
View Daniele-Cangi's full-sized avatar
🎯
Focusing
🎯
Focusing
286 followers · 252 following

Achievements

Achievement: Pair ExtraordinaireAchievement: Pull Sharkx2Achievement: QuickdrawAchievement: YOLO

Achievements

Achievement: Pair ExtraordinaireAchievement: Pull Sharkx2Achievement: QuickdrawAchievement: YOLO

Block or report Daniele-Cangi

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
Daniele-Cangi/README.md
Systems fail where assumptions begin.

Independent Systems Researcher & Builder

Build → Measure → Attack assumptions → Verify → Restrict claims → Harden

Current work

XCP Studio

An experimental source-to-target adaptation system using Xbox Series X as a real constrained execution target.

XCP began as a deterministic compute worker and evolved into a broader system for source observation, Typed Semantic IR, constrained target execution, differential analysis, causal attribution and evidence-bound equivalence.

Development is currently private; selected research artifacts and reproducible evidence may be released separately.

WorldLoop

A procedural CAD and world-generation system for Roblox.

WorldLoop evolved from an architecture generator into a CAD-first system for creating editable buildings, terrain and procedural environments, with deterministic delivery into Roblox Studio.

Other product work

Co-author — LLegal

Legal evidence and state infrastructure built around Danish legal sources, combining authority-aware retrieval, temporal validity, document analysis and provenance.

Co-author — Roomee

A Denmark-focused roommate and room-matching platform with server-authoritative matching, privacy controls and mobile product infrastructure.

The common thread

The domains change, but the method is consistent: build the system, observe where the model stops matching reality, make the assumptions explicit, strengthen the evidence, and change the architecture when necessary.

Some projects are developed independently and others through UnityLoop.

Hugging Face · LinkedIn

Pinned Loading

  1. Aurora Aurora Public

    C++20 research system for adaptive FEC, cross-layer control, and resilient heterogeneous networking under energy and link constraints.

    C++ 5 1

  2. XCP-Research XCP-Research Public

    Public evidence for XCP Studio — a creative software system for AI-assisted creation, source-to-target adaptation, and measured execution on Xbox Series hardware.

    Python 2

  3. Alchimista Alchimista Public

    Self-hosted AI governance, document intelligence, RAG, privacy and auditable decision evidence — with local-first deployment and optional Full Rizzo PII detection.

    Python 4 2

  4. Derivative Derivative Public

    Software synthesis engine that turns requirements into verified artifacts — or explicit failure evidence — through execution-grounded validation.

    Python 4

  5. SolidWindow SolidWindow Public

    Decision engine for flexible electricity demand: finds cheap power windows, validates whether they can be trusted, and audits decisions through replay.

    Python 3 3

  6. Neuro-Trace Neuro-Trace Public

    Transformer interpretability research framework for circuit discovery, sparse features, and controlled residual-stream interventions.

    Python 3