Experience

Riverain Technologies

• United States
• Medical Device
• 1 - 100 Employee

• Principal Research Scientist

  • Feb 2023 - Present

  Ohio, United States Developing new techniques for detection and prediction of lung cancer in CT scans



SRI

• United States
• Research Services
• 700 & Above Employee

• Advanced Computer Scientist

  • Sep 2014 - Nov 2022

  Princeton, New Jersey Key Projects: - DARPA COMBAT: Developing an AI for army combat training simulator (OneSAF) using deep reinforcement learning to automatically generate multi-agent behavior trees. - DARPA SAIL-ON: Developing AI architectures that are capable of Lifelong Learning and can detect and respond to unexpected, novel situations and environments. - DARPA QED-MEDIQ: Low bit-precision Deep Learning networks for speed and robustness against adversarial attacks. - Work with LatentAI: Developing… Show more Key Projects: - DARPA COMBAT: Developing an AI for army combat training simulator (OneSAF) using deep reinforcement learning to automatically generate multi-agent behavior trees. - DARPA SAIL-ON: Developing AI architectures that are capable of Lifelong Learning and can detect and respond to unexpected, novel situations and environments. - DARPA QED-MEDIQ: Low bit-precision Deep Learning networks for speed and robustness against adversarial attacks. - Work with LatentAI: Developing bit-compressed classification and object detection networks for power-efficient edge computing. - DARPA NGA-BIG: Designing and training semantic segmentation and automatic Scene Graph generation for complex entities in Satellite Imagery. - TSWG-GRIFN / ONR-WAR: GPS-denied geolocation and heading estimation using semantic segmentation and skyline feature matching. - DARPA XAI-DARE: Deep Attentional methods for Explainable AI. - FHWA DCODE: Deep learning for upper body pose estimation of vehicle drivers. - AR Stryker: Augmented reality for armored vehicles. Other internal R&D projects: - Developing deep predictive coding networks for real-time depth estimation; - Developing novel Neural Inertial Navigation algorithms (navigation using IMU only without visual odometry). - Developing auditory processing networks for identifying speech/utterances of autistic children during therapy sessions to quantitatively measure progress. - Investigating potential applications of neuro-inspired deep learning algorithms utilizing event cameras for power-efficient and robust visual odometry. Show less



New York University

• Higher Education
• 700 & Above Employee

• Postdoctoral Research Scientist

  • Oct 2013 - Aug 2014

  New York City Psychophysics on Deep Learning models: Measured thresholds in noise of human observers and Convolutional Neural Networks on the same visual tasks to bridge the fields of psychophysics and machine learning. Identified the network architectural constraints and training paradigms under which CNNs exhibit several characteristic hallmarks of human vision extensively studied in the psychophysics literature: robustness to position uncertainty, feature grouping, crowding and spatial frequency… Show more Psychophysics on Deep Learning models: Measured thresholds in noise of human observers and Convolutional Neural Networks on the same visual tasks to bridge the fields of psychophysics and machine learning. Identified the network architectural constraints and training paradigms under which CNNs exhibit several characteristic hallmarks of human vision extensively studied in the psychophysics literature: robustness to position uncertainty, feature grouping, crowding and spatial frequency channels. Show less



Columbia University

• United States
• Higher Education
• 700 & Above Employee

• PhD Candidate

  • Jan 2008 - May 2013

  New York City Computational neuroscience data analysis and modeling: Quantified the degree of diversity in tuning of neurons to help constrain developmental models of visual cortical circuitry. Also developed an algorithm to reduce errors in classifying spikes from different neurons (“spike sorting”)



Weizmann Institute of Science

• Israel
• Research
• 700 & Above Employee

• Summer Intern

  • Jun 2007 - Aug 2007

  Rehovot, Israel Modeling stochastic cellular systems (protein trafficking, DNA repair): Characterized behavior of polymers under different stochastic regimes and estimated (mathematically and using numerical simulations) the average time it would take for a DNA fragment to find a small target in the chromatin structure or for mRNA to find a small target.



Princeton University

• United States
• Higher Education
• 700 & Above Employee

• Undergraduate Researcher

  • May 2006 - Jun 2007

  Princeton, NJ Studying visual information coding principles in the retina: Measured consistency of responses of retinal ganglion cells responding to the same stimulus and determined the synergy and redundancy of information between sets of retinal ganglion cells. Developed a theoretical framework to analyze how heterogeneity of neural firing rates in a large population of neurons affects coding efficiency and ability to discriminate between inputs.



Princeton University

• United States
• Higher Education
• 700 & Above Employee

• Undergraduate Researcher

  • May 2005 - Jan 2006

  Princeton, NJ Studied the use of beating electrostatic waves as a novel, energy-efficient method for heating plasma for rocket propulsion. Characterized the conditions under which such heating occurs.



Department of Ophthalmology, Tygerberg Hospital

• Summer Intern

  • Jul 2004 - Aug 2004

  Cape Town, South Africa Developed Java software to facilitate clinical tests of an application of stochastic resonance to clinical ophthalmology