MAFFINTECH

Cortex

AI hedge-fund infrastructure for emerging markets

A production ML quant engine. Every trade decision carries a calibrated probability, a confidence interval, and an explanation — end-to-end automated, with no human in the loop.

Built by MAFFinTech for emerging-market brokers.

Detect Decide Execute

Book a briefing See how it works

July 2026 · Cairo, Egypt

THE PROBLEM

Emerging markets have data, not infrastructure.

RETAIL

Has data. Lacks alpha.

MENA retail traders have access to feeds and apps — but no regime-aware ML, no calibrated probabilities, no institutional execution layer.

BROKERS

Want products. Can't build the stack.

Emerging-market brokers see customer demand for AI-managed funds — but a quant ML stack is 24–36 months of engineering they don't have.

REGULATORS

Opened the door in 2024.

Egypt's FRA introduced the Robo-Advisor License (Decree 57/2024). A clear path exists. No production-grade platform exists to fill it.

THE PRODUCT

Cortex — production ML quant engine.

End-to-end automated. No human in the loop. Multi-asset by design.

50+

structure detectors & features

proprietary multi-timeframe market-structure features

ML models & optimizers

regime · scoring · sizing — architecture shared under NDA

humans in the loop

detect → decide → execute, fully automated

100%

explainable decisions

calibrated probability + a logged explanation on every call

Explore the full pipeline

VALIDATION

Six months live on the EGX100.

H1 2026 — every trade decision tracked in real time.

9,345

EGX100 trades evaluatedJan–Jun 2026 · live

49.8%

day-trading win ratepowers the automation · break-even is 29.9% — a 20-pt cushion

66.7%

position-trading win rateoffered to users by their broker · break-even is 38.3% — a 28-pt cushion

+32.5%

H1 returnposition trading · 1M EGP · 3% per trade · gross of costs

METHODOLOGY

Two EGX100 strategies ran through H1 2026 with every trade decision — accepted or rejected — tracked live with identical exit logic. Day-trading selection powers the automation; position-trading selection is the product a broker offers its clients. H2 tracking is already running, with every new decision flagged live.

See the full H1 2026 results & charts

ENGINEERED FOR EMERGING MARKETS

We don't bolt a US model onto a new market.

Cortex flexes to each market's microstructure and rules. Our first localization is the Egyptian Exchange.

EGYPT

FRA

Robo-Advisor License

BoD Decree 57/2024. Defined pathway with EGP 15M (~$300K) capital requirement.

Our primary market

UAE

ADGM FSRA

Category 3A · Retail

Most permissive global fintech regime. Covers discretionary portfolio management.

Expansion target — 2027

SAUDI ARABIA

CMA

QFI restrictions lifted

Tadawul retail market opening to foreign participation Feb 2026. No robo decree yet.

Expansion target — 2028

EGX localization & go-to-market

FOUNDERS

Three founders. Ten years of collaboration.

The founders run every briefing personally — methodology questions welcome.

Mohamed Fekry

CEO · Technical Founder

Built the Cortex ML pipeline personally. Trading infrastructure since 2020.

Islam Alla

CTO · Technical Founder

Data analytics and Python engineer. Full-time on MAFFinTech since June 2023.

Alexandra Mereuta

CMO · BD & Partnerships

Originated the company's anchor commercial partnership and leads BD. Multi-lingual operator. Full-time since mid-2025.

A stack that conventionally takes 24–36 months — built by a tiny team with heavy AI-assisted development under founder supervision, plus 2 operational hires. No external contractors.

See what Cortex can do for your brokerage.

A 30-minute technical briefing — the methodology, the validation, and how a partner goes live.

Book a briefing