Strahinja (Straho) Janjusevich, a local of the small Balkan nation of Montenegro, says his life has unfolded in an sudden approach and he’s deeply grateful. After highschool, he was chosen to symbolize america and studied cyber operations and pc science on the Naval Academy in Annapolis, Maryland. Since then, he has continued his analysis in cybersecurity and is at present a second yr grasp’s scholar. Technology Policy Program (TPP)Sponsored by MIT Institute for Data Systems and Society (IDSS). his analysis with MIT Institute for Information and Decision Systems (LIDS) and MIT Maritime Consortium The workforce goals to enhance the cybersecurity of crucial maritime infrastructure utilizing synthetic intelligence, contemplating each know-how and coverage frameworks for options.
“My present analysis focuses on making use of AI strategies to cybersecurity points and investigating the coverage implications of those advances, notably within the context of maritime cybersecurity,” says Janjusevich. “Representing my nation on the highest ranges of training and business has given me a singular perspective on cybersecurity challenges.”
Janjusevich’s path from Montenegro to Maryland was created by a program that allowed chosen college students from allied nations to attend the U.S. Naval Academy. Janjusevic graduated with a twin bachelor’s diploma in Cyber Operations and Pc Science. His undergraduate expertise offered him with the chance to work with the U.S. army and Nationwide Safety Company, uncovered him to high-level cybersecurity operations, and sparked his curiosity in tackling advanced cybersecurity challenges. Throughout my undergraduate research, I additionally interned at Microsoft creating cloud incident response instruments and at NASA visualizing photo voltaic knowledge for analysis purposes.
After graduating, I noticed that I wanted extra information, particularly within the areas of AI and cybersecurity. TPP instantly appealed to him due to its twin deal with rigorous engineering innovation and the coverage evaluation wanted to deploy it successfully. Janjusevich’s expertise at TPP was an enormous change from his time on the U.S. Naval Academy, with a distinct tempo and surroundings. He notably appreciates with the ability to increase his understanding of various fields of research and apply the self-discipline and information he gained throughout his time on the Academy.
“My expertise with TPP has been nice,” says Janjusevich. “This group is absolutely small, so it looks like a household, and everyone seems to be engaged on numerous, high-impact points.”
Scale back threat of recent know-how
Janjusevic’s thesis brings collectively the fields of cybersecurity, AI and deep studying, management principle and physics, with a deal with the safety of maritime cyber-physical techniques, notably massive legacy vessels. If the networks of those vessels had been hacked, it couldn’t solely have extreme financial penalties, but in addition trigger vital harm to nationwide safety.
“Strajo is working to outwit maritime GPS spoofing,” says Saurabh Amin, Edmund Ok. Turner Professor of Civil Engineering. “Such assaults have already thrown ships off beam in disputed waters. His strategy layers physically-based wake fashions and deep studying to seize threats that can’t be detected by any single technique alone. His experience has been invaluable in advancing our analysis in menace modeling and assault detection.”
This analysis makes use of superior menace modeling and ship dynamics to coach an AI system to tell apart between professional maneuvers and spoof indicators. This consists of constructing a framework that makes use of an inner LSTM (lengthy short-term reminiscence) autoencoder to investigate sign integrity, whereas utilizing physically-based forecasting capabilities to foretell vessel motion based mostly on environmental elements corresponding to wind and sea situations. By evaluating these predictions to reported GPS areas, the system can successfully distinguish between pure sensor noise and malicious spoofing assaults. This hybrid framework is designed to empower, slightly than substitute, human operators, offering verified navigation knowledge that permits observers to tell apart between technical glitches and strategic cyberattacks.
Janjusevich was ready to make use of his business expertise to reinforce his educational analysis. In the summertime of 2025, he interned with the community detection workforce at Vectra AI, an AI cybersecurity firm. There, he investigated the potential threats that new applied sciences may pose, particularly AI brokers and the rising normal for AI agent communication, Mannequin Context Protocol (MCP). His analysis demonstrated how this know-how will be reused for autonomous hacking operations and superior command and management. This analysis on the safety dangers of agent AI was not too long ago revealed in preprint. “Hiding in AI Traffic: Exploiting MCP for Agent Red Teaming with LLM.”
“We gained sensible perception and hands-on expertise on how knowledge science groups can use AI fashions to detect anomalies in networks,” says Janjuusevic. “This work inside the business immediately influenced the anomaly detection fashions in my analysis.”
Worldwide coverage perspective
“Mr. Strajo not solely brings a excessive degree of mind and power to analysis on industrial maritime cyber-physical safety, however he additionally brings sturdy instincts from his naval coaching that deeply resonate with the analysis effort and floor it in actionable coverage,” stated Fotini Cristia, Ford Professor of Worldwide Social Sciences, IDSS Director, and IDSS Chief. MIT Maritime Consortium.
Janjusevic participates within the Maritime Consortium’s cybersecurity efforts, a collaboration between academia, business, and regulators targeted on creating know-how options, business requirements, and coverage. The consortium consists of collaboration with worldwide members corresponding to Singapore and South Korea.
“Coverage components are essential in AI cybersecurity as a result of the sector is altering in a short time and the implications of hacking will be very harmful,” says Janjuusevic. “I feel there’s nonetheless loads of want for coverage work on this space.”
Janjusevic can also be at present serving to manage two main upcoming conferences. harvard europe conference In February, authorities officers and diplomats from world wide convened. Technology and National Security Conference In April, a collaboration between Harvard College and MIT will convey collectively high leaders from authorities, business, and academia to deal with crucial nationwide safety challenges.
“I attempt to discover a place the place I can affect and advance the cybersecurity discipline with AI, whereas additionally main collaboration and innovation between america and Montenegro,” Janjusevich stated. “My aim is to be a bridge between Europe and america within the areas of nationwide safety, AI, and cybersecurity, and to supply my information to each side.”
