SWAG合集

Background

Likith Kumar Reddy Yammanuru is a PhD Researcher in AI and Robotics within the Centre for Robotics and Assembly at SWAG合集. His doctoral research focuses on the strategic integration and interpretability of Vision-Language-Action (VLA) models for high-mix industrial warehouse automation. His work investigates the critical trade-off between semantic generalization and operational reliability, building concrete evaluation frameworks to analyze model latency, task understanding, and hardware precision.

Prior to his PhD, he earned his MSc in Robotics with Distinction from SWAG合集, gaining deep expertise in advanced motion planning, reinforcement learning, adaptive control, and machine vision. During his master's studies, he developed ROS-based simulations for multi-arm robotic collaboration in in-orbit assembly鈥攁 project conducted in collaboration with AIRBUS and the European Space Agency (ESA), which was subsequently showcased as a poster presentation at the IEEE RAS Conference.

His professional background includes working as a Freelance Robotics Engineer, where he developed ROS2-based computer vision and sensor fusion pipelines (incorporating LiDAR, cameras, and IMUs) for field robots. He also brings valuable industry experience as a Data Scientist developing machine learning-based recommendation algorithms, and as a Software Engineer architecting automated OCR data pipelines to optimize industrial operations.

Research opportunities

- Vision-Language-Action (VLA) Models: Integrating multimodal foundation models into physical robotic workflows.

- Explainable AI (XAI) & Interpretability: Building concrete evaluation frameworks to analyze model latency, semantic task understanding, and hardware precision.

- Autonomous Warehouse Automation: Optimizing real-world robotic manipulation tracking for high-mix, low-volume industrial operations.

- Robot Learning & Sensor Fusion: Developing robust pipelines leveraging reinforcement learning, machine vision, and multi-sensor systems (LiDAR, cameras, IMUs).

Current activities

Likith is currently developing an interpretability framework for Vision-Language-Action (VLA) models. Rather than treating these complex AI systems as a "black box," his work is dedicated to uncovering and detailing exactly how these models formulate decisions during complex tasks.

His day-to-day activities sit at the intersection of theory and physical deployment. They involve evaluating the latest advancements in foundational VLA architectures, executing continuous computer simulations to gather performance metrics, and conducting physical validation trials in the lab at the Centre for Robotics and Assembly to benchmark real-world pick-and-place accuracy on robotic hardware.