My research develops advanced data-driven AI solutions at the intersection of Trustworthy AI, healthcare, and information security. I design secure, decentralized decision-support systems using multimodal data integration, federated learning, and neuro-symbolic AI. A key focus is on neuro-symbolic architectures and knowledge-graph representations for interpretable reasoning and robust predictive analytics.
In cybersecurity, I build AI-enabled forensic pipelines for verifying the authenticity, integrity, and veracity of digital media, including deepfake detection, leveraging multimodal feature fusion, representation learning, and explainable AI. Collaborations with industry focus on deploying neuro-symbolic AI for applications such as web filtering and threat detection.
In healthcare, I apply multimodal AI and predictive modeling to forecast cerebrovascular and cardiovascular events by integrating clinical text with multiple imaging modalities (e.g., DSA, MRA) using deep learning, attention mechanisms, and graph-based representations to improve clinical decision support.
Please describe one or two of your most interesting projects.
I develop an AI detector for forged media, including deepfake images, videos, and audio. The system combines multimodal analysis and neuro-symbolic reasoning to deliver robust, explainable, and interpretable forensic insights.
I also lead VascularInsightAI, a federated, multimodal AI tool that predicts cardiovascular and cerebrovascular events, including aortic and cerebral aneurysm rupture leading to subarachnoid hemorrhage. By integrating clinical text, lab data, and medical imaging, it provides scalable, privacy-preserving, and interpretable risk predictions for critical healthcare decisions.
How did you end up where you are today? (Your research journey)
My journey into AI research began with a fascination for creating systems that are not only intelligent but also trustworthy and interpretable, particularly in areas where mistakes can have serious consequences. Early in my career, I recognized that combining data-driven learning with structured reasoning could bridge the gap between predictive performance and explainability. This led me to explore neuro-symbolic AI and knowledge graphs, developing architectures that integrate deep learning with structured knowledge to produce robust, transparent, and reliable models.
Over time, my focus expanded to critical, high-stakes domains. In healthcare, I apply these methods to predict cardiovascular and cerebrovascular events, including aneurysm rupture and subarachnoid hemorrhage, by combining clinical text, lab data, and imaging to provide actionable, privacy-preserving risk assessments. In cybersecurity, finance, and judicial contexts, I build AI tools to detect forged media and deepfakes, providing interpretable forensic insights to help prevent decisions based on manipulated or fraudulent information.
Along the way, I have been guided by the goal of bridging methodological innovation with real-world impact, designing AI systems that are not only accurate but also ethical, explainable, and deployable in practice. My journey has been one of continually pushing the boundaries of AI while keeping its applications meaningful and responsible.
What is the most significant scientific contribution you would like to make?
I aim to develop truly trustworthy AI systems that combine deep learning with neuro-symbolic reasoning and knowledge graphs to deliver robust, interpretable, and privacy-preserving decisions. My work focuses on creating AI that can explain its reasoning, detect inconsistencies, and adapt safely in high-stakes domains such as healthcare, cybersecurity, finance, and the courts. The ultimate goal is to establish AI that people and institutions can reliably trust, bridging the gap between cutting-edge machine learning and responsible, human-aligned decision support.
What makes you excited about your data science and AI research?
What excites me most about my research is building trustworthy, human-centered AI that can tackle high-stakes problems in healthcare, cybersecurity, finance, and the courts. I focus on neuro-symbolic, multimodal AI that not only predicts accurately but also explains its reasoning and collaborates effectively with humans. I’m equally passionate about mentoring students and developing curricula that train professionals to work responsibly and effectively with AI.
What are 1-3 interesting facts about yourself?
I lead interdisciplinary AI research integrating neuro-symbolic methods to enhance AI-human teaming, mentor students, and develop curricula that empower professionals to work with AI.