His research focuses on the intersection of efficient deep learning systems and algorithms. His work "On-device training under 256KB memory" (MCUNetV3) is highlighted by the MIT homepage in fall 2022. Ligeng's projects have been integrated into frameworks such as PyTorch and AutoGluon, and have been covered by media outlets including MIT News and IEEE Spectrum. He was awarded the Qualcomm Innovation Fellowship, and his research has received over 3,100 citations on Google Scholar and over 8,000 stars on GitHub.