Renbin Liu
About Renbin Liu
Renbin Liu is a Machine Learning Scientist at Flexport with a background in data analysis, engineering, and research, holding degrees from MIT.
Current Title at Flexport
Renbin Liu is currently working as a Machine Learning Scientist at Flexport in Bellevue, Washington, United States. This hybrid role sees Liu applying machine learning methodologies to optimize logistics and supply chain solutions for the company.
Educational Background
Renbin Liu studied Computer Science and Engineering at the Massachusetts Institute of Technology (MIT). He earned a Bachelor of Science (BS) from 2017 to 2021 and a Master of Engineering (MEng) with a focus on Artificial Intelligence (AI) from 2021 to 2022. Liu's master's thesis was notably recognized with an award from MIT.
Previous Experience at MIT
Renbin Liu has extensive research and teaching experience at the Massachusetts Institute of Technology. He worked as an Undergraduate Research Assistant from 2020 to 2021, a Lab Assistant in 2020, an Undergraduate Researcher from 2019 to 2020, and an Undergraduate Teaching Assistant in 2019 and 2021. He also served as a Graduate Teaching Assistant in 2022. This blend of academic roles provided Liu with a robust foundation in both research and mentorship.
Internships and Early Career Roles
Renbin Liu's early career includes multiple internships. He was a Data Analyst Intern at Blendoor in San Francisco and the San Francisco Bay Area in 2021 and 2020. Liu also interned as a Data Engineer at Fidelity Investments in Durham, North Carolina, in 2020 and a Data Science Intern at Sky in London, England, in 2021. These positions allowed him to gain hands-on experience in data analysis and engineering.
Significant Project Contributions
Renbin Liu made substantial contributions to the development of Milestone Prediction Service (MPS) for ocean vessels. He improved model accuracy by 10 percentage points through data quality enhancement and feature augmentation. Liu also accelerated modeling work by creating an automated orchestration workflow, saving approximately four engineering weeks. Additionally, he designed a gRPC API endpoint to reduce milestone latency significantly, increasing the coverage of historical events from 60% to 92%.