Marios
Savvides
Marios
Savvides
Marios
Savvides
Founder of CMU’s Center for Foundational Intelligence, Marios Savvides holds an endowed professorship and a research record spanning
80+ patents, 250+ publications, and global-leading citations.
His breakthroughs in long-range iris, unconstrained face, and real-time vision systems have moved from academic research into deployed biometric infrastructure for the Department of Defense,
Homeland Security, and major international airports.
Founder of CMU’s Center for Foundational Intelligence, Marios Savvides holds an endowed professorship and a research record spanning 80+ patents, 250+ publications, and global-leading citations.
His breakthroughs in long-range iris, unconstrained face, and real-time vision systems have moved from academic research into deployed biometric infrastructure for the Department of Defense, Homeland Security, and major international airports.
Founder of CMU’s Center for Foundational Intelligence, Marios Savvides holds an endowed professorship and a research record spanning 80+ patents, 250+ publications, and global-leading citations.
His breakthroughs in long-range iris, unconstrained face, and real-time vision systems have moved from academic research into deployed biometric infrastructure for the Department of Defense, Homeland Security, and major international airports.
His lab’s work also powers large-scale commercial systems, including autonomous inventory robots in 550 Walmart stores and the high-performance video-analytics platform that became the foundation of Oosto’s $100M+ ARR business. Beyond research, he has launched companies like HawXeye and Ultron AI, driving AI technologies into millions of devices and multiple venture-backed outcomes.
Elected in the National Academy of Inventor's 2025, honorary recipient of the "Outstanding Contributor in AI" Award by the U.S. Secretary of the Army, and named "Inventor of the Year Award" in 2022 by the Pittsburgh Intellectual Property Law Association (PIPLA), Savvides leads one of the most commercially and operationally impactful AI labs in the country.
Latest interviews
World Economic Forum
Advanced Facial Recognition
Technology
Advanced Facial Recognition Technology
NOVA Science Channel
The Boston Manhunt: What Technologies Worked and Which Didn’t
The Boston Manhunt: What Technologies Worked and Which Didn’t
CBS 60 Minutes
Interview with Leslie Stahl,
“Faces in a crowd”
Interview with Leslie Stahl, “Faces in a crowd”
CNN
Facial Recognition Failed in Boston Attack
Advanced Facial Recognition Technology
CNN
Iris Eye Scanners Getting Longer Range
Iris Eye Scanners Getting Longer Range
In the press
01_08_2021
Popular Science
“Facial recognition still can’t prove someone’s identity”
07_29_2020
CNN Business
“Face Masks Are Screwing Up Facial Recognition Software”
10_16_2018
Popular Science
“Get ready for a lot more facial recognition at the airport”
01_17_2018
Carnegie Mellon
“CMU Experts Lead Conversations at World Economic Forum”
“CMU Experts Lead Conversations at World Economic …”
09_29_2017
The Hill
"iPhone X’s facial recognition is cool, but it’s not the future of technology"
"iPhone X’s facial recognition is cool, but it’s not the …"
09_21_2016
PBS
“How Artificial Intelligence Can Stop Sex Trafficking”
11_26_2013
CNN
"Embracing big brother: How facial recognition could help fight crime"
"Embracing big brother: How facial recognition …"
07_28_2009
Scientific American
Scientific Ameri…
"Security on the (Eye)Ball: Hands-Free Iris Biometrics to Keep Bad Guys …"
"Security on the (Eye)Ball: Hands-Free Iris Biometrics …"
01_08_2021
Popular Science
“Facial recognition still can’t prove someone’s identity”
07_29_2020
CNN Business
“Face Masks Are Screwing Up Facial Recognition Software”
10_16_2018
Popular Science
"Get ready for a lot more facial recognition at the airport"
01_17_2018
Carnegie Mellon
"CMU Experts Lead Conversations at World Economic Forum - News - Carnegie Mellon University"
09_21_2016
PBS
"How Artificial Intelligence Can Stop Sex Trafficking"
09_29_2017
The Hill
"iPhone X’s facial recognition is cool, but it’s not the future of technology"
11_25_2013
CNN
"Embracing big brother: How facial recognition could help fight crime"
07_28_2009
Scientific American
"Security on the (Eye)Ball: Hands-Free Iris Biometrics to Keep Bad Guys at Bay"
The Center for Foundational Intelligence accelerates the scientific foundations of AI. Our researchers collaborate across disciplines to advance theory, architect new models, and build systems that expand what intelligent machines can achieve. Through this work, CFI provides the essential groundwork that fuels future breakthroughs and supports impactful applications today.
Learn more
Learn more
Industry experience
2023 - Present
Founded UltronAI, developing edge-optimized visual foundation models for retail loss prevention and real-time product understanding. Leading work in ultra-low-power inference (2–4W) on Hailo and NVIDIA Jetson, scalable SKU-level recognition, and energy-aware embedded AI. Built on 80+ patents from his CMU lab to address the $170B global shrink and staffing problem.
2020 - 2023
Served as Chief AI Scientist at Oosto (formerly AnyVision), advancing real-time computer-vision platforms deployed across critical-infrastructure, school-safety, and industrial-security environments. Contributed to the AI systems behind what became a $100M+ ARR analytics platform.
2017 - 2020
Served as Chief AI Scientist at Bossa Nova Robotics, where he and his CMU lab re-architected the entire visual-AI stack for autonomous inventory robots. Scaled deployments from 20 to 550 Walmart stores in 18 months, running multiple daily missions fully autonomously and eliminating human-in-the-loop operations.
2014 - 2018
Co-founded and served as CTO of HawXeye Inc., developing low-form-factor embedded AI for legacy home-security cameras, enabling robust movement and object detection. Resulted in over 6 million cameras shipped with Savvides-led AI and culminated in a successful acquisition.
2004 - Present
Founded and directs the CMU Biometrics Center (re-launched in 2025 as the Center for Foundational Intelligence). Over two decades, his lab has produced 80+ patents, 250+ publications, and foundational breakthroughs in iris recognition, face detection, and embedded AI for national security, aviation, retail, and consumer devices.
Founded UltronAI, developing edge-optimized visual foundation models for retail loss prevention and real-time product understanding. Leading work in ultra-low-power inference (2–4W) on Hailo and NVIDIA Jetson, scalable SKU-level recognition, and energy-aware embedded AI. Built on 80+ patents from his CMU lab to address the $170B global shrink and staffing problem.
2023 - Present
Served as Chief AI Scientist at Oosto (formerly AnyVision), advancing real-time computer-vision platforms deployed across critical-infrastructure, school-safety, and industrial-security environments. Contributed to the AI systems behind what became a $100M+ ARR analytics platform.
2020 - 2023
Served as Chief AI Scientist at Bossa Nova Robotics, where he and his CMU lab re-architected the entire visual-AI stack for autonomous inventory robots. Scaled deployments from 20 to 550 Walmart stores in 18 months, running multiple daily missions fully autonomously and eliminating human-in-the-loop operations.
2017 - 2020
Co-founded and served as CTO of HawXeye Inc., developing low-form-factor embedded AI for legacy home-security cameras, enabling robust movement and object detection. Resulted in over 6 million cameras shipped with Savvides-led AI and culminated in a successful acquisition.
2014 - 2018
Founded and directs the CMU Biometrics Center (re-launched in 2025 as the Center for Foundational Intelligence). Over two decades, his lab has produced 80+ patents, 250+ publications, and foundational breakthroughs in iris recognition, face detection, and embedded AI for national security, aviation, retail, and consumer devices.
2004 - Present
Notable Awards
NAI
"National Academy of Inventors" Award
Professor Marios Savvides is named a 2025 Fellow of the National Academy of Inventors in recognition of his outstanding contributions to artificial intelligence and patented innovation.
Professor Marios Savvides is named a 2025 Fellow of the National Academy of Inventors in recognition of his outstanding contributions to artificial intelligence and patented innovation.
U.S. ARMY
Outstanding Contributor in AI
Professor Marios Savvides receives the "Outstanding Contributor in AI" in area of advanced face recognition and detection by the U.S. Secretary of Army, Mr. Ryan McCarthy, March 2020.
Professor Marios Savvides receives the "Outstanding Contributor in AI" in area of advanced face recognition and detection by the U.S. Secretary of Army, Mr. Ryan McCarthy, March 2020.
PIPLA 2022
Inventor of the Year Award
Professor Marios Savvides named the "Inventor of the Year Award "by the Pittsburgh Intellectual Property Law Association, PIPLA. 2022.
Professor Marios Savvides named the "Inventor of the Year Award "by the Pittsburgh Intellectual Property Law Association, PIPLA. 2022.
CARNEGIE MELLON
Steven J. Fenves Award for Systems Research
Professor Marios Savvides receives the "Stevens J. Fenves Award for Systems Research" from Dean of College of Engineering, Bill Sanders, Carnegie Mellon University, 2023.
Professor Marios Savvides receives the "Stevens J. Fenves Award for Systems Research" from Dean of College of Engineering, Bill Sanders, Carnegie Mellon University, 2023.
2020 EDISON AWARDS
1st Place - "Autonomous Data Capture and Analysis of On-Shelf Inventory"
We received the 1st Place Gold Award in 2020 Edison Awards for the Artificial Intelligence based Retail Video Analytics we developed.
We received the 1st Place Gold Award in 2020 Edison Awards for the Artificial Intelligence based Retail Video Analytics we developed.
2020 ARTIFICIAL INTELLIGENCE EXCELLENCE AWARD
Theory of Mind
Professor Savvides receives the "Theory of Mind" Award in 2020 for his Artificial Intelligence research.
Professor Savvides receives the "Theory of Mind" Award in 2020 for his Artificial Intelligence research.
GLOBAL PITTSBURGH
Immigrant Entrepreneur Award in Technological Innovation
Professor Savvides receives Global Pittsburgh Immigrant Entrepreneur Award in Technological Innovation in Dec 2018.
Professor Savvides receives Global Pittsburgh Immigrant Entrepreneur Award in Technological Innovation in Dec 2018.
CARNEGIE MELLON
The Bossa Nova Robotics Artificial Intelligence Endowment
Presented by CMU President Farnam Jahanian to Professor Savvides, Bossa Nova Robotics CEO and CMU Board of Trustees Dr. Bruce McWilliams, CMU Provost Jim Garrett in November 2018.
Presented by CMU President Farnam Jahanian to Professor Savvides, Bossa Nova Robotics CEO and CMU Board of Trustees Dr. Bruce McWilliams, CMU Provost Jim Garrett in November 2018.
Trusted partners
Citations: 16,200 | h-index: 66 | i10-index: 207
2025
MetaVLA: Unified Meta Co-training For Efficient Embodied Adaption
2025
KnowledgeSmith: Uncovering Knowledge Updating in LLMs with Model Editing and Unlearning
KnowledgeSmith: Uncovering Knowledge Updating in LLMs with …
2025
STELAR-VISION: Self-Topology-Aware Efficient Learning for Aligned Reasoning in Vision
STELAR-VISION: Self-Topology-Aware Efficient Learning for Aligned …
2025
SOLAR: Scalable Optimization of Large-scale Architecture for Reasoning
2025
Rtgen: Generating region-text pairs for open-vocabulary object detection
2025
A Reference-Based 3D Semantic-Aware Framework for Accurate Local Facial Attribute Editing
A Reference-Based 3D Semantic-Aware Framework for Accurate Local …
2025
Powering finetuning in few-shot learning: Domain-agnostic bias reduction with selected sampling
Powering finetuning in few-shot learning: Domain-agnostic bias reduction…
Displaying 5 of 16,200
2025
System and method for multiview product detection and recognition
2025
System and method for shrinkage detection and prevention in self-checkout
2025
Two-way product verification by reverse lookup using multi-view enrollments
2025
Single shot 3D modeling from 2D image
2025
Two-way product verification by reverse lookup
Displaying 5 of 72
Citations: 16,200 | h-index: 66 | i10-index: 207
2025
MetaVLA: Unified Meta Co-training For Efficient Embodied Adaption
2025
KnowledgeSmith: Uncovering Knowledge Updating in LLMs with Model Editing and Unlearning
2025
STELAR-VISION: Self-Topology-Aware Efficient Learning for Aligned Reasoning in Vision
2025
SOLAR: Scalable Optimization of Large-scale Architecture for Reasoning
2025
Rtgen: Generating region-text pairs for open-vocabulary object detection
2025
A Reference-Based 3D Semantic-Aware Framework for Accurate Local Facial Attribute Editing
2025
Powering finetuning in few-shot learning: Domain-agnostic bias reduction with selected sampling
Displaying 5 of 16,200
2025
System and method for multiview product detection and recognition
2025
System and method for shrinkage detection and prevention in self-checkout
2025
Two-way product verification by reverse lookup using multi-view enrollments
2025
Single shot 3D modeling from 2D image
2025
Two-way product verification by reverse lookup
Displaying 5 of 72