I am a Ph.D. candidate at the University of Tennessee, Knoxville, currently focusing on large-scale representation learning for poultry action recognition at the UT Smart Agriculture Lab. My research emphasizes building and benchmarking models that perform reliably in real-world agricultural environments, leading to projects such as ChickenVerse, an open multi-task dataset for chicken detection, segmentation, and behavior recognition, and ChickenSense, a low-cost deep learning solution for acoustic poultry feed monitoring.
Before my current research, my background was in robotics and autonomous systems, where I developed software for humanoid soccer robots competing in the RoboCup Humanoid League. Our team achieved 1st Place in 2015 (China) and 3rd Place in 2014 (Brazil), an experience that sharpened my skills in real-time perception, embedded programming, and vision-based decision making and produced my early open-source work: HumanoidSoccerRobot and Pro-IMU.
I am also deeply interested in NLP for low-resource languages, particularly Persian, and passionate about open-source tools and datasets that strengthen the Persian AI ecosystem, most notably Shekar, a Python toolkit for Persian NLP, and NeyShekar, a large-scale open Persian speech dataset.