Embodied intelligence
Computational approaches to intelligence grounded in perception, action, and interaction with the physical world.
Building Cognitive Systems | Computer Vision & Embodied Intelligence
Intelligence that perceives and acts in real time.
I have spent nearly four decades bridging fundamental AI research, industrial R&D, entrepreneurship, product development, and university teaching—always asking how intelligent systems can understand and interact with the physical world.
4MIT degrees
6issued U.S. patents
€5mEU research programs originated or co-authored
Multimilliondollars in DARPA research funding won and led
Research & engineering vision
From scientific questions to complete systems.
My work asks how intelligent systems acquire information from the world, extract meaningful structure, and use that understanding to guide behavior under real operational constraints.
Real-time perception
Complete perception-to-action systems designed for operational responsiveness in unconstrained environments.
Research translation
Fundamental work in AI, vision, and statistical modeling carried into defense and commercial systems.
Selected experience
Research, teaching, and systems development.
Academic inquiry, funded research leadership, entrepreneurship, and products that had to work reliably in the real world.
2023–present
Koç University
Instructor, Computer Science and Artificial Intelligence
Teach undergraduate computer science and a graduate course in Embodied Artificial Intelligence developed from the ground up, while supervising capstone work in computer vision, machine learning, and intelligent systems. Repeatedly recognized among the university’s top 10% of instructors.
2017–2019
Autometrics
Chief Science Officer
Personally designed and developed an end-to-end statistical system for measuring the effect of individual television advertisements on consumer vehicle-purchase intent using large-scale demographic and market data.
2010–2016
Co57 Systems
Founder & Chief Science Officer
Won and led a four-year, multimillion-dollar DARPA program in automated visual understanding of human activities. Set the scientific direction and system architecture for biologically inspired real-time perception systems.
2006–2010
MIT CSAIL
Research Scientist
Originated two successful DARPA research seedlings, including an $800,000 program on visuospatial reasoning whose results helped form the basis for DARPA’s MindsEye program.
2001–2004
COGVIS EU Project
Research Scientist · LIRA-Lab, University of Genova
Conceived a three-stage developmental architecture for robotic cognition and implemented core components of its first stage—work singled out by EU reviewers as outstanding and novel.
1997–1998
Imagen
Co-Founder
Co-founded an early image-based search venture with two fellow MIT AI Laboratory graduate students. The concept won the Grand Prize in the 1997 MIT $50K Entrepreneurship Competition.
1992–1993
Xerox PARC
Research Intern · Two summers
Developed Sketchlyzer, a system for interacting with computers, fax machines, and copiers through hand-drawn sketches. The research produced six issued U.S. patents.
Teaching & mentorship
Connecting foundational ideas to technical practice.
At Koç University, I teach undergraduate computer science and a graduate course in Embodied Artificial Intelligence that I developed from the ground up.
I draw on academic research, industrial R&D, and entrepreneurship to connect foundational concepts with real-world examples, while supervising capstone work in computer vision, machine learning, and intelligent systems.
CurrentlyInstructor
Department of Computer Engineering
Koç University, Istanbul
Repeatedly recognized, based on student course evaluations, among the university’s top 10% of instructors.
Selected publications & patents
Published research and patented inventions.
Work spanning active vision, developmental cognition, learning through action, and computational models of attention.
- Learning about Objects through Action: Initial Steps toward Artificial CognitionIEEE International Conference on Robotics and Automation, 2003
- Learning to Act on ObjectsBiologically Motivated Computer Vision, Springer-Verlag, 2002
- Inducing the Focus of Attention by Observing Patterns in SpaceIJCAI Workshop on Modeling Others from Observations, 2005
- Repairing Learned Knowledge Using ExperienceArtificial Intelligence at MIT: Expanding Frontiers, MIT Press, 1990
Invited talks
Communicating ideas across disciplines.
Selected recent talks for research, policy, industry, and public audiences.
The Influence of Brain and Cognitive Science on My Work in AI
Koç University Neuroscience Festival
Demystifying AI: What Innovators Need to Know
Tufts University Gordon Institute
Constructing Intelligence
Norway Financial Summit
A Model of Visual Attention for Event Perception and Recognition
University of Coimbra
Education
Four degrees from MIT.
PhD · Artificial Intelligence
Dissertation: Visual Routines and Attention
SM · Electrical Engineering and Computer Science
Thesis: Repairing Learned Knowledge Using Experience
SB · Brain and Cognitive Science
SB · Mathematics with Computer Science (Course 18-C)
Recognition
Selected honors.
Recognized among Koç University’s top 10% of instructors
Invited to seven European Commission expert review cycles evaluating research proposals in robotics and AI
Selected as a Kauffman Fellow
Grand Prize, MIT $50K Entrepreneurship Competition
MIT Peter J. Eloranta Undergraduate Research Award
Elected Full Member, Sigma Xi, The Scientific Research Honor Society.
