Apar Pokhrel

Hello! I’m Apar Pokhrel, a Research Assistant at the UT Arlington Research Institute working on Computer Vision and Autonomous Systems to enable intelligent pereception and decision-making. I also conduct research in Neuromorphic Vision and Machine Learning, advised by Dr. Diego Patiño.

My research sits at the intersection of Computer Vision, Embodied AI, and Machine Learning, with the goal of building intelligent systems that can perceive, understand, and interact with the physical world.

Currently, my work spans three main areas:

Computer Vision & Foundation Models

I’m interested in building visual learning systems that can understand the world from images, video, and multimodal data. My research explores ViTs, VLMs, and self-supervised learning to develop scalable visual representations that generalize across tasks with minimal supervision. More broadly, I study how foundation models can improve perception, reasoning, and visual understanding for downstream applications.

Neuromorphic Vision

My work investigates event-based cameras as an alternative sensing paradigm for high-speed and high-dynamic-range perception. I study self-supervised representation learning, temporal modeling, and multimodal fusion techniques that combine asynchronous event streams with conventional RGB imagery. I’m particularly interested in enabling robust perception under challenging conditions where traditional frame-based cameras struggle.

Embodied AI & Robotics

I develop perception systems for autonomous robots and UAVs, ranging from real-time edge deployment on NVIDIA Jetson platforms to Vision-Language-Action models that enable robots to understand and execute complex tasks. Recently, I’ve been working towards VLM models to investigate how perception, language, and action can be unified to build agent in real-world environments.