Skip to content

DisuzaQuantitative/Disuza-Quantitative

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Disuza Quantitative

Systematic Crypto Trading Research Laboratory

Madrid, Spain · Founded 2025 · Version 3

Python Google Cloud CC BY 4.0 Wikidata Last Commit

Living technical reference — architecture, methodology, and regulatory posture. Source code and model weights remain proprietary.

Website · Guest Portal · LinkedIn · LLM Context · FAQ


Non-solicitation notice. This repository is informational documentation only. Disuza Quantitative does not offer regulated investment services or solicit retail or professional investor capital. No statement in this repository constitutes an offer, invitation, or inducement to engage in any investment activity in any jurisdiction. Statements about regulatory alignment are expressions of internal posture, not representations of regulatory approval or licensed status. Before engaging with Disuza Quantitative in any capacity, consult local counsel.


What Disuza is

Disuza Quantitative is a private quantitative trading research laboratory engineering systematic execution algorithms for digital asset perpetual futures (BTC, ETH) and US equity index futures (NQ, ES). The engine is a rule-based systematic engine with a layered risk overlay that adjusts exposure in response to short-horizon regime shifts; machine learning is restricted to auxiliary roles (meta-labeling, dynamic position sizing, regime detection) per López de Prado's framing — never as the primary signal generator. The platform runs on Google Cloud and executes through prop-trading-programme APIs and self-custody perpetual venues under trade-only permissions.

What Disuza is NOT

  • Not a high-frequency trading firm.
  • Not a retail-facing signal-selling service.
  • Not a public hedge fund — access is closed and by invitation.
  • Not licensed at this time — operating in a pre-licensing phase, monitoring the MiCA and FINMA regulatory landscapes as part of our pre-licensing preparation. No statement herein constitutes a representation of regulatory compliance or approval.

Architecture snapshot

graph LR
  A[Multi-source ingestion<br/>exchange data · venue WS · macro · on-chain regime] -->|PIT pipelines| B[Feature store]
  B --> C[Rule-based signal engine<br/>+ layered risk overlay]
  C --> D[Signal router]
  D -->|Pub/Sub| E[Execution]
  E --> F[Prop-trading-programme APIs]
  E --> G[Self-custody perpetual venues]
  D -.->|broker-truth reconciliation| F
  D -.->|broker-truth reconciliation| G
  style A fill:#f3f3f3,stroke:#333
  style C fill:#e8f4f8,stroke:#333
  style E fill:#f0e8f8,stroke:#333
Loading

Full diagrams live in docs/diagrams/. Architecture walkthrough in docs/architecture.md.

Explore the documentation

Section What you'll find
Overview What Disuza is, the thesis, the regulatory posture
Architecture Data plane · compute plane · execution plane · persistence
Data pipeline Multi-source ingestion with point-in-time guarantees
Execution Venue classes, non-custodial execution, broker-truth reconciliation
Risk framework Tiered drawdown posture, non-custodial guarantees, audit trail
Regulatory MiCA and FINMA landscape, pre-licensing posture, disclosures
Technology Language, frameworks, orchestration, cloud, observability
Components The services that make up the production platform
Operations Deployment, monitoring, alerting, reliability posture
Team Yasmine Bendhiab and Fares Bendhiab
Skills Technical competencies applied to the platform
FAQ Questions about access, technology, regulation, and collaboration

Team

Yasmine Bendhiab — Co-Founder & CEO · Strategic Operations & Corporate Compliance Madrid, Spain · LinkedIn

Fares Bendhiab — Co-Founder & CTO · Lead Architect of the Quantitative Infrastructure Bizerte, Tunisia · LinkedIn · GitHub

Regulatory snapshot

  • Jurisdiction: Kingdom of Spain, registered entity in Madrid.
  • Phase: Pre-licensing.
  • Monitoring: MiCA (EU Markets in Crypto-Assets Regulation) and FINMA (Swiss Financial Market Supervisory Authority) regulatory landscapes.
  • Simulated performance: CFTC Rule 4.41 acknowledged. Any backtest results referenced on the public guest portal are hypothetical historical simulations.
  • Custody model: Non-custodial — credentials scoped to trade-only operations across both execution classes.
  • Access model: Closed-access, invitation-only. Not soliciting retail investors.

Detail in docs/regulatory.md.

Contact

All inquiries (general, partnerships, careers): contact@disuza.com

Documentation licence

This documentation is released under Creative Commons Attribution 4.0 International (CC BY 4.0) with explicit carve-outs for trademarks, source code, model weights, and operational know-how. See LICENSE for the full text and the no-endorsement clause required for adaptations.


Disuza Quantitative is a closed-access research laboratory. This repository contains architecture and methodology notes as a living public reference. Source code, model weights, and operational know-how are proprietary.

© 2025-2026 Disuza Quantitative · CC BY 4.0