Amman ↔ BudapestCorvinus University of BudapestUpdated 09 Aug 2026

Research / Code / Evidence

I build empirical systems that have to survive contact with the data.

I’m Layan Aloreidi, a computational empirical researcher building causal evidence, real-time economic data systems and open research software across financial markets, AI and labor, and MENA economies.

See the flagship work ↓Download CV GitHub

01DOI-backed research software

14real-time forecast models

31Terminal tests at alpha release

2026–27frozen prospective holdout

Three flagship research programs

Not a project inventory. A body of evidence.

Each program exposes its methods, limitations and current validation state.

Award winnerPublic softwareProspective validation active

Financial networks × information theory × research software

From award-winning paper to falsifiable software.

The Emergent Market Mind models markets as changing information networks. MarketMind MII translates that framework into tested Python software and freezes the regime-to-signal hypothesis before a year-long confirmatory holdout.

Research contribution

Market Intelligence Index combining memory, information flow and network connectivity.

Engineering proof

Typed API, CLI, dashboard, causal normalization, cost-aware walk-forward evaluation and deterministic demo.

Open-science proof

PyPI distribution, Zenodo DOI, frozen OSF registration, deviation log and public validation ledger.

Claim boundary

Historical results generate the hypothesis; they are not presented as prospective confirmation.

Repository Install from PyPI Zenodo DOI Frozen validation

MII

MEMORY / .35FLOW / .40NETWORK / .25

Public research benchmarkFrozen protocol

Macroeconomics × real-time data × forecast evaluation

Fourteen models. Vintage information. No revised-data time travel.

A brutal forecasting tournament that asks when complexity actually earns its keep. Models are evaluated at rolling origins using the information available then—not the data economists can see years later.

MODEL FAMILYREAL-TIME TESTOUTPUT

Naive / ARIMA / VAR / BVARALFRED vintagesPoint + density

Factor / state-spaceRolling originCalibration

Regularized / tree boostingRegime slicesDM comparison

Open tournament Technical bench

14default modelsplus naive baselines

Executable alphaRights-aware release gate

MENA data × provenance × historical vintages

What could a researcher have known on that date?

The MENA Economic Terminal is data infrastructure built around releases, vintages, revisions, source artifacts and exact transformations. It refuses to silently replace history with the latest revised number.

menaecon / vintage query

mena.get(
  "unemployment",
  country="JOR",
  vintage="2026-06-01"
)

Research-integrity decision: the first official Jordan transcription remains quarantined while reuse rights and an external source-document audit are unresolved.

Repository and alpha MENA Open Data & Evidence Lab

01release_timerecorded

02vintagerecorded

03revisionrecorded

04source_hashrecorded

05transformationrecorded

06git_commitrecorded

Proof ledger

Every claim gets a maturity level.

No stars, downloads or private praise are counted as validation. Public work moves up only when evidence changes.

ObjectCurrent evidenceNext gate

MarketMind MIIAward + DOI + tests + preregistrationIndependent reproduction

Macro tournamentSealed real-time protocol + public artifactsExternal rerun

MENA TerminalExecutable alpha + 31 tests + exact captureRights clearance + source audit

Econometrics libraryPublic simulations + CI + failure casesClean-environment validation

MENA AI × LaborAudited seed + acquisition frame50-firm pilot + second coder

Open external validation tasks

Reproduce MarketMind Rerun macro leaderboard Audit source capture Test notebook install Second-code AI labels

Working research + methods

The next proof, not the next pile of features.

Flagship economics paper / data constructionPilot in progress

MENA Firm-Level AI Adoption & Labor-Market Adjustment

Building a provenance-traced 2018–2025 listed-firm panel to distinguish generic AI language from substantive operational deployment and study labor adjustment.

01Audited real-data seed

0220-firm acquisition frame

0350-firm validation pilot

04Second coder + agreement

05Freeze, then estimate

Inspect the public build

Modern econometricsValidation phase

Econometrics Notebook Library

Estimands, identification logic, simulations, implementations and reproducible failure cases for modern causal methods.

Sun–Abraham, Callaway–Sant’Anna and BJS imputation
Wild-cluster bootstrap and Conley HAC
Shift-share, synthetic control, randomization inference and local projections

Read the notebooks

Open economic evidenceFounder-produced pilot

MENA Narrative & Market Stress Observatory

A small provenance-first pilot with a data dictionary, processing code, tests, corrections policy and an explicit independent-rerun gate.

Audit the pilot

Open-source contributionOpen / awaiting re-review

PyFixest PR #1456

Documentation for fitted-model result accessors. This is presented as an open contribution—not a merged upstream claim.

View upstream PR

Economics / decision theoryApproved for publication

Too Smart to Wait

A theory of endogenous timing reversal: stronger AI capability can intensify competitive pressure to act before information is socially mature.

Positive Science — Economics · 2026

Research collaboration

Give me a question that cannot be answered because the dataset does not exist.

Available for empirical research assistance, reproducibility work, data construction, research-software collaboration and MENA economic evidence projects.

Email GitHub ORCID LinkedIn Research CV