DART Academy

Online scams are a rapidly growing threat, particularly for older adults. In a single year, 201,266 Americans aged 60+ reported being victims of online scams, resulting in $7.7 billion in losses, according to the FBI's Annual Internet Crime Report.

Reporting by The New York Times has revealed that many of these scams are not isolated efforts by individual fraudsters, but highly organized, industrial-scale operations run by transnational crime syndicates. They run detailed internal "playbooks" that train workers step by step in how to manipulate victims and defraud them of their money. If criminal networks rely on structured training to carry out fraud at scale, people need equally structured education to recognize and resist these tactics.

To address this problem, a multi-university research collaboration led by the University at Buffalo launched DART Collective (Deception Awareness and Resilience Training), a nonprofit supported by a $5 million National Science Foundation Convergence Accelerator grant. The project brings together researchers from institutions including Cornell University, Lehigh University, Clemson University, Northeastern University, and the University of Illinois to develop tools that help people recognize and resist online deception.

As a technical lead and founding engineer, I worked with researchers and educators across the collaboration to design and build DART Academy, an interactive learning system that teaches users to identify scams through realistic simulations, structured lessons, and hands-on practice. I led product design and prototyping in Figma and contributed to NSF National I-Corps customer discovery interviews, translating insights from 109 interviews into the platform’s user flows and features.

I architected and built the system's learning management infrastructure, including the course player, content authoring tools, and interactive training framework. I also assembled and led a rotating development team of student engineers over multiple years, mentoring developers through system design, implementation, and deployment as the platform evolved.

A course authoring system for cybersecurity and scam awareness training that enables educators to easily create interactive lessons. Built as a custom content management system, it supports continuous content creation so learners receive timely, relevant, and engaging training on emerging scams. A block-based interface makes it possible to build rich end-to-end learning experiences that combine content, interaction, and assessment.

An AI-powered voice generation system that transforms course content into natural, instructor-ready narration. It generates speaker notes from lesson content, prepares them for voice output, and integrates directly into the editor for quick review and refinement. The system uses the OpenAI API for content generation and AWS Polly for voice synthesis, with a focus on producing clear and engaging narration that presents material in an accessible way for older adult learners. This streamlines the content creation process while enhancing the learning experience with high-quality audio narration.

A modular course delivery system designed to guide learners smoothly through structured lessons and real-world scam scenarios. Built as a single-page application with preloaded content and seamless fading transitions, it reduces friction and creates a continuous learning experience. A sidebar supports self-paced navigation across chapters and lessons, while a presentation mode with voiceover narration allows content to be delivered in guided sessions. This helps learners stay engaged and better recognize scam patterns through practice.

Learners practice scam recognition through realistic simulations. Each module mirrors common fraud patterns and gives immediate feedback on choices.

The Alert Hub ensures learners are exposed to the latest scam tactics as they evolve, reinforcing training with timely, real-world context. The system automatically updates with relevant news, with the ability for teams to review and refine content as needed to maintain accuracy and quality.

A multilingual support system with locale-aware routing for English, Spanish, French, and Chinese, designed to scale to additional languages. It includes internationalization (enabling locale-based routing and language support without code changes) and localization (adapting content, metadata, interface text, and URLs for each language) to ensure accessibility across diverse user populations. The system is production-ready, with translated content currently in development.

A custom donor management system built as an alternative to third-party donation platforms, reducing overhead while providing greater flexibility and control. Contributions are kept in sync across the platform database, Stripe, and HubSpot for reliable tracking and reporting. Four giving tiers structure the experience: Supporter ($10), Contributor ($25), Advocate ($50), and Champion ($100), each tied to a specific impact statement. Donors can contribute through one-time, monthly, or annual giving, with support for custom amounts. The system includes donor-only access to exclusive downloads such as a top scam tips checklist, stickers, and printable course PDFs for offline learning, along with early releases of new courses and profile badge recognition. An admin layer provides tools for managing donor status, tier progression, and engagement analytics.

A gamification system that drives learner engagement through points, rank progression, and leaderboard visibility. Learners earn points for completing lessons, quizzes, and practice activities, progressing through rank tiers from Novice to Legendary as they build skills over time. Leaderboards highlight top performers while showing each learner’s relative position, encouraging continued participation. Progress tracking gives learners a clear sense of advancement across lessons and courses. Immediate feedback reinforces learning in real time, while completion certificates mark key milestones and can be downloaded and printed as a tangible record of achievement.