Projects

What we have built, and what it does now.

We assemble specialists from Harvard, MIT and Stanford around each problem, then build what the research demands.

Project study 1

Future-Proof Open Source Intelligence (OSINT) Early Warning System

Partner
Harvard University
Sector
Critical Infrastructure
Public Sector
Advisors
Prof. Bruce Huang PhD Ed.D
Dr. Stephen Elston PhD
Research team
5 Harvard Data Science Candidates
Context

Critical infrastructure operators and public sector organisations face risks that surface in open information long before they appear in internal reporting. The signals exist — in global news, public datasets and search behaviour — but they arrive faster and in greater volume than any analyst team can read, and conventional reporting cycles surface them only after the fact.

National Innovation Labs established this initiative with Harvard University to address that gap: an industry-led project developing a future-proof open source intelligence early warning system for critical infrastructure and public sector organisations globally.

Our services

We assembled a research team of Harvard Data Science candidates, guided by advisors Professor Bruce Huang PhD Ed.D and Dr. Stephen Elston PhD, and took the problem from research question to working system.

The team designed and built an ingestion layer drawing on GDELT and Google Trends, applied natural language processing — TF-IDF, named entity recognition and BERTopic — to identify and track emerging themes, and layered STL decomposition and Z-score models over the result to detect anomalies against expected patterns. Findings surface through Apache Superset, on infrastructure built in Python on AWS.

This system was built using:

AWS — S3, Lambda Python TF-IDF NER BERTopic STL decomposition Z-score models GDELT Google Trends Apache Superset
Outcome

The result is a running early warning capability rather than a report: emerging risk themes detected from open sources, anomalies flagged against a modelled baseline, and reporting an analyst team operates itself.

The system is designed to be future-proof — new data sources and detection models can be added without rebuilding the pipeline — and it is architected to run inside a client's own security boundary. Development continues with our Harvard research partners.

OSINT early warning system dashboard

Open the live dashboard

Research advisors

The initiative is guided by two doctoral advisors.

Professor Bruce Huang PhD Ed.D

Professor Bruce Huang PhD Ed.D

Bruce is a technology executive, academic leader and researcher at the intersection of AI, data science, technology strategy and responsible innovation. He leads Harvard Extension School's Master's Program in Information Technology and brings extensive industry and academic experience spanning IBM, entrepreneurship and applied technology research.

Specialises in
AI & Machine Learning AI Governance, Ethics & Responsible AI Data Science & Analytics Technology Strategy & Digital Transformation Data Engineering & Information Systems Innovation & Emerging Technologies Technology & Organisational Leadership
Dr. Stephen Elston PhD

Dr. Stephen Elston PhD

Stephen is a data scientist, machine learning engineer and AI specialist with decades of experience applying advanced analytics to complex business problems. He has taught AI, machine learning and data science at Harvard Extension School, the University of Washington and edX, and advises organisations from global enterprises to startups.

Specialises in
AI & Machine Learning ML Engineering & Predictive Analytics Data Science, Analytics & Visualisation Real-Time & Streaming Analytics Financial Analytics & Risk Modelling Fraud Detection & Electronic Payments Data & Cloud Architecture Machine Vision & Signal Processing
Project team

The researchers below are the team delivering this project, drawn from our research bench and matched to the problem.

Priyanka Gunduboina

Priyanka Gunduboina

Master of Liberal Arts (ALM) — Data Science Candidate

GIS Analyst at Capitol Airspace Group with a dual B.S. in Environmental Policy & Planning and Biological Sciences from Virginia Tech. Her background spans AWS, renewable energy analytics, and geospatial data systems across government and private-sector organizations. She specializes in spatial analysis, data management, and environmental risk modeling.

Sergey Onokhin

Sergey Onokhin

Master of Liberal Arts (ALM) — Data Science Candidate

Senior Pricing Analyst in BP's Supply, Trading & Shipping organization, driving pricing strategy and margin optimization for North American B2B fuels. U.S. Army Infantry veteran with a background in mission-critical operations.

Michael Cabrera

Michael Cabrera

Master of Liberal Arts (ALM) — Data Science Candidate

Research Data Analyst for the State of California with a background in Electrical Engineering from UC San Diego. Served as a U.S. Patent Examiner at the USPTO, specializing in computer science and human-machine interface technologies. His experience spans technical evaluation, data analysis, and systems-level review across complex software and innovation domains.

Connor Poetzinger

Connor Poetzinger

Master of Liberal Arts (ALM) — Data Science Candidate

Production-focused data scientist at OKX combining AI, machine learning and Python expertise to solve complex business and financial crime challenges.

Phu Thai

Phu Thai

Master of Liberal Arts (ALM) — Data Science Candidate

Software Engineer at Microsoft within Azure Software Defined Networking, focusing on cloud networking infrastructure that connects global datacenter systems. He holds a B.S. in Computer & Systems Engineering from Rensselaer Polytechnic Institute and brings deep expertise in networking, distributed systems, and cloud architecture.

Tell us what needs to exist. We'll tell you what it takes to build it.

Contact us