Researcher · Founder

Abdullah X.

I research how advanced AI systems behave, and how institutions can govern them.

Founder and President of Project AWARE, an international think tank focused on institutional AI safety.

Selected research
Research

Selected publications

Technical AI safety, institutional design and governance.

Empirical AI safety, model behavior and institutional design.

Selected preprints, working papers and manuscripts. Scholarly citations and other documented references are distinguished in the source panels below. Source register reviewed 15 November 2026.

Unlearning at Scale: State-Exact Trace-Preserving Deletion Machine learning
4 references

Documented sources (4)

Additional coverage (not counted)

2025 ↗
Projects

Selected projects

Research tools, experiments, and working systems.

AI SAFETY POLIS An experimental platform for designing and testing institutions for multi-agent AI systems. AI EVALUATION SafeguardShift Measuring agent behavior when verification, review, and other operational safeguards are lost. ML RELIABILITY PipelineProof Execution-based verification of silent integrity failures in machine-learning pipelines. EVALUATION SCIENCE Probe Scent Testing how readily AI models recognize scenarios that resemble evaluations. MACHINE UNLEARNING Unlearning at Scale Trace-preserving training replay for reproducible, state-exact data deletion. INTERPRETABILITY Markovian Circuit Tracing Testing transformer state abstractions against known dynamics and causal ground truth. SOCIAL SIMULATION LLM Persona Studying when better synthetic individuals fail to produce more faithful populations. ALGORITHMIC FAIRNESS Status Association Audit Separating socioeconomic associations from their effects on consequential AI decisions. TELECOMMUNICATIONS ICR-OPT Simulating mobile-network sharing, capacity constraints, and cooperative settlements. SIGNAL PROCESSING OpenVQ A speech-quality research engine with interpretable signal and telecom diagnostics. OPERATIONS RESEARCH Queue Topology Lab Interactive simulation and optimization of finite-capacity service networks. COMPUTATIONAL SYSTEMS Path-Dependent Radix Spaces Exact ranking, enumeration, and deterministic sampling of structured finite domains. FINANCIAL COMPUTING FinSpace Rank-addressable financial scenarios for reproducible testing and simulation. PUBLIC GOVERNANCE Public Sector AI Survey A paired-choice study of public preferences for oversight of AI-assisted decisions. EDUCATION DATA UDISE+ Exposure Study Reproducible data pipelines for school access, composition, and infrastructure research. ECONOMIC RESEARCH SEC Cyber Disclosure Timing Analyzing cybersecurity disclosure timing and financial-market responses.
A selection of public research and engineering work. All repositories

Attention.

Engagements

Selected engagements

Public lectures, expert panels and convenings.

Experience

Work & appointments

Selected roles, past and present.

Work experience

10 roles
Founder & PresidentProject AWARENov 2024 to Present
Founder & CEOZephara AIAug 2023 to Present
Research AssociateAI Governance Taskforce, Arcadia ImpactNov 2026 to Present
Guest FacultyIILM University2025 to Present
Applied Policy FellowPolicy Agora, ORF & Esya CentreJul to Aug 2026
Policy-in-Action FellowYoung Leaders for Active CitizenshipOct 2025 to Feb 2026
AI EngineerCompliDataFeb to Oct 2023
ConsultantCompliDataJul 2022 to Feb 2023
Consultant180 Degrees ConsultingFeb to May 2023
Board Member & Acting CFOUndivided vzw, Belgium2021 to 2022

Appointments & memberships

3 appointments
Member, Global Working GroupMAP-AI Global Working Group on Safe & Trusted AI2026 to Present
Member, National Working Group on AIGovernment of India2026 to Present
Member, CoRE-AIThe Dialogue2024 to Present
Background

A wider perspective

Research and institution-building across disciplines.

I founded Project AWARE to study practical ways of governing increasingly capable AI systems. My work brings technical experiments into conversation with questions of institutional authority, public policy and international cooperation.

EDUCATION

Aligarh Muslim UniversityEconomics
Takshashila InstitutionPublic Policy
KU LeuvenElectronics & ICT coursework
Contact

A good conversation
starts here.

For research, institutional collaborations or anything in between.