Applied research for prediction, security, and civic-scale intelligence.
Building systems that detect signals, explain change, and support better decisions.
Our programs analyze patterns, predicts emerging trends, identifies risks, and make complex data more usable.
Areas of interest
A practical research agenda organized around prediction, trust, security, and access.
-
Applied machine learning
Designing ML systems that move beyond demos: classification, clustering, recommendations, model evaluation, and workflows that support real users.
-
Trend and prediction analysis
Exploring how signals emerge over time through forecasting, anomaly detection, comparative metrics, dashboards, and explainable trend models.
-
Cybersecurity research
Studying authentication, verification, abuse prevention, infrastructure hardening, incident patterns, and defensive automation.
-
Educational AI
Investigating how intelligent learning tools can support instructors, students, feedback cycles, assessment, and personalized educational pathways.
-
Public data systems
Building browsable, searchable, and machine-readable access layers for public documents, regulatory filings, civic data, and government records.
-
Infrastructure resilience
Researching reliability across software, Linux systems, automation pipelines, data ingestion, reproducible builds, and fault-tolerant deployment practices.
Research principles
Practical
Research should produce working systems, not only abstract models.
Explainable
Predictions and classifications should be traceable, auditable, and understandable.
Public-interest oriented
Technical systems should improve access, accountability, resilience, and community benefit.
Project pipeline
Collect
Gather structured and unstructured data from public records, educational systems, technical logs, APIs, and research datasets.
Model
Apply machine learning, statistical analysis, natural language processing, and forecasting techniques to identify useful signals.
Explain
Build interfaces, summaries, dashboards, and documentation that make model behavior and findings understandable.
Deploy
Package research into usable tools, reproducible code, public demos, articles, and infrastructure that can be tested and reused.