Greater Seattle Area
Applied scientist, tech lead and software engineer specializing in multimodal LLMs (MLLMs), video foundation models, computer vision, search and robotics. Experienced in driving high-impact AI solutions across Amazon, Qualcomm, and GE. Led cross-team collaborations to enable research-to-product transitions of LLM/MLLM systems, video foundation models, and sensor fusion stacks for autonomous driving. Delivered state of the art (SOTA) results in CVPR/ECCV publications and 19 granted US patents.
Tech lead for successful AI product launches across multiple teams - Launched an agentic search tool to identify relevant video clips for marketing campaigns, and an automated agentic workflow to identify content ratings for videos, at Amazon Prime Video - Launched a finetuned MLLM to identify and display helpful customer review images in Amazon Rufus the AI shopping assistant - Pre-trained and fine tuned state of the art video foundational models for video retrieval and classification powering product video and ad video retrieval in Amazon product search - Launched object tracking algorithms for a fleet of sidewalk delivery robots at Amazon Scout Mentored research projects on video models published at CVPR, ACL and ECCV
Algorithm design and software development for camera-radar sensor fusion and vehicle tracking in autonomous driving. Radar beamforming design. Deep learning for camera and radar perception. Lidar processing.
Developed data-driven and physics-based statistical inference, signal processing, and data analysis algorithms for failure prediction and reliability modeling in industrial machines such as power generators, wind turbines, cutting tools and CT machines. Contributed to research grant applications to government agencies. Played a major role in securing a $5.3 M grant on insider threat detection from Intelligence Advanced Research Projects Activity (IARPA), jointly with the University at Albany, SUNY and GeorgiaTech, 2015.
Research included topics in signal processing and information privacy. Particularly focused on two problems: signal processing, sampling theory, and trajectory design for data acquisition with moving sensors, and the quantification and protection of privacy in data collection and information processing. Taught Master's level courses in Signal Processing and supervised Master's and PhD student research projects
Conducted research in the fields of Detection and estimation, Information theory and Wireless communications. Primarily focused on problems in statistical decision-making, exploring both the theoretical aspects of optimal statistical tests, as well as practical applications to spectrum sensing and dynamic spectrum access for cognitive radio systems.
Teaching assistant for the graduate level course on "Optimization by Vector Space Methods", offered by Prof.Tamer Basar.