Strahinja Janjusevic, a native of Montenegro, has forged an extraordinary path from his homeland to the forefront of maritime cybersecurity research at MIT. After being selected to represent Montenegro at the U.S. Naval Academy in Annapolis, he earned dual bachelor’s degrees in cyber operations and computer science. His time at the academy exposed him to elite cybersecurity operations through collaborations with the U.S. military and the National Security Agency, while internships at Microsoft and NASA further sharpened his technical skills in cloud security and data visualization. Driven by a desire to deepen his expertise, Janjusevic joined MIT’s Technology and Policy Program (TPP), hosted by the Institute for Data, Systems, and Society (IDSS). There, he combines engineering rigor with policy insight, focusing on securing maritime cyber-physical systems—especially large, aging commercial vessels vulnerable to cyberattacks. His research, conducted with the MIT Laboratory for Information and Decision Systems (LIDS) and the MIT Maritime Consortium, centers on countering GPS spoofing, a growing threat that can mislead ships in contested waters, endangering both national security and global trade. Janjusevic’s approach integrates artificial intelligence with physics-based modeling. He developed a hybrid system that uses a deep learning LSTM autoencoder to analyze GPS signal integrity, while simultaneously employing a physics-driven forecaster to predict a vessel’s expected trajectory based on environmental conditions like wind and sea state. By comparing real-time GPS data with these predictions, the system can detect subtle spoofing attempts that traditional methods might miss. The goal is not to replace human operators but to provide them with trusted, verified navigation data to distinguish between technical errors and deliberate cyberattacks. His work has been recognized by Saurabh Amin, the Edmund K. Turner Professor in Civil Engineering, who praised Janjusevic’s ability to bridge AI and physical system dynamics to improve threat detection. In summer 2025, Janjusevic interned at Vectra AI, a leading AI cybersecurity firm, where he explored emerging risks in agentic AI, particularly the Model Context Protocol (MCP), a new standard for AI agent communication. His findings revealed how MCP could be exploited for autonomous hacking and advanced command-and-control operations. This research was published in the preprint “Hiding in the AI Traffic: Abusing MCP for LLM-Powered Agentic Red Teaming,” highlighting the urgent need for security safeguards in next-generation AI systems. Janjusevic also brings a strong international perspective to his work. As a member of the MIT Maritime Consortium, he contributes to efforts that unite academia, industry, and regulators—including partners from Singapore and South Korea—to shape standards and policies for maritime cybersecurity. He emphasizes that rapid technological change demands equally agile policy responses, especially in high-stakes domains like shipping and national security. Beyond research, Janjusevic is actively shaping discourse in the field, helping organize major conferences such as the Harvard European Conference and the Technology and National Security Conference, which bring together global leaders to address critical security challenges. His ultimate goal is to serve as a bridge between Europe and the United States, leveraging his Montenegrin roots and American training to advance cybersecurity innovation, policy, and international collaboration in the age of AI.