Omega Makena Mwebia
Independent researcher and architect of Scarcity.
Who I Am
A Quantitative Researcher and Machine Learning Engineer. I build mathematical models and software architectures that discover structure in chaos.
What I Do
I research and develop intelligent systems capable of learning invariant relationships under extreme data scarcity and non-stationarity.
How I Do It
Through rigorous quantitative analysis and the development of specialized frameworks—primarily Scarcity—that push compute directly to the mathematical boundaries of the problem.
Why I Do It
To understand complex systems. I believe that by modeling what systems retain when they cannot retain everything, we uncover their fundamental organizational identity.
Scarcity
Scarcity is a structure-discovery framework for evolving systems. Explained in simple terms, Scarcity learns the organizational structure of dynamic data and converts it into reusable graph priors for anomaly detection, forecasting, and downstream learning under limited-data conditions.
Explore ScarcityResearch Notes
View Archive →Scarcity: A Streaming Relationship-Discovery and Federated-Learning Framework for Multivariate Time Series Under Data Scarcity
Preprint outlining the Scarcity framework, its methodology, and its measured behaviour.
Scarcity Benchmark 06: Statistical Calibration and Weakness Audit
Synthetic ground-truth validation, typed discovery evaluation, weakness audits, and formal statistical calibration pipeline.
Scarcity Benchmark 05: Relationship Structure Discovery
Advanced relationship structure discovery, real-data scarcity verdicts, and comprehensive harness details.
Scarcity Benchmark 04: Robustness, Reproducibility, and Limitations
Reproducibility harness, stress tests, failure modes, privacy analysis, and final claim integrity.
Scarcity Benchmark 03: Economic Simulation and Ablations
Validation of directional economic simulation, confidence anchoring, and structural ablation studies.
Scarcity Benchmark 02: Core Results and Data Scarcity
Main prediction and discovery results, non-IID verification, and data scarcity curves.
Scarcity Applications
View Full Network →KShield
Agent-based scenario testing across national sectors to evaluate policy shocks safely.
KPulse
Real-time social sensing and early-warning narrative intelligence algorithms.
KCollab
Interagency unified command spaces for synchronized tactical responses.
KScarcity
The unified national dashboard tracking real-time risks and interventions.