Researcher
CurrentAs part of my previous laboratory team, I applied graph neural network and reinforcement learning to hardware logic synthesis for EDA tools, improving efficiency. I designed GNN models to predict circuit delays and constructed state-representations for RL agents. My work included new machine learning approaches for logic synthesis and novel applications of RL in graph structures.Skills: Machine Learning, Neural Network Design, Computer Vision, Reinforcement Learning, Graph Neural Network, Logic Synthesis Optimization.Responsibilities:- Improved compressed model accuracy by 20% with symmetric quantization.- Developed flexible run-time bit-width adaptation for object detection.- Conducted research on RL, binarization/quantization, and GNN state representation.- Extracted data from OpenSTA and analyzed it using Cppyy.- Utilized Pytorch packages for ML algorithm development and visualization.- Proficient in requirement gathering, analysis, estimation, testing, and reporting.