Qin L.
About Qin L.
Qin L. is a Staff Machine Learning Research Engineer at Scale AI, where he leads the development of the Scale RLXF distributed training platform and has a strong background in machine learning techniques. He has previously worked at Microsoft and Baidu Inc., and holds multiple degrees in computational science and machine learning from prestigious universities.
Work at ScaleAI
Currently, Qin L. serves as a Staff Machine Learning Research Engineer at Scale AI. Since joining in 2023, Qin has led the development of the Scale RLXF distributed training platform. This platform emphasizes large language model (LLM) and multimodal post-training techniques, including supervised fine-tuning (SFT), preference modeling, and reinforcement learning from human feedback (RLHF) and alignment. Qin employs various methods such as Proximal Policy Optimization (PPO), Direct Preference Optimization (DPO), Knowledge Transfer Optimization (KTO), and rejection sampling in these initiatives. Additionally, Qin developed Scale's LLM and data auto-train-evaluation pipeline, which is widely used for data operations and model customization tasks.
Previous Experience at Microsoft
Before joining Scale AI, Qin L. worked at Microsoft as a Senior Research Software Development Engineer (SDE) from 2017 to 2022. During this five-year tenure in Bellevue, Washington, Qin contributed to various projects, focusing on advancements in machine learning and software development. This role allowed Qin to gain significant experience in research and development within a leading technology company.
Experience at Baidu Inc.
Qin L. also held the position of Senior Research Developer at Baidu Inc. from 2014 to 2016. This two-year role in Beijing involved working on innovative projects in machine learning and artificial intelligence, contributing to Baidu's research initiatives and product development efforts.
Education and Expertise
Qin L. has an extensive educational background in computational science and engineering, as well as machine learning. Qin earned a Master of Science in Computational Science and Engineering from Harvard University, studying there from 2016 to 2017. Prior to that, Qin completed a Master of Science in Machine Learning at Peking University from 2011 to 2014. Qin also holds a Bachelor of Engineering in Automation and Mathematics from Tsinghua University, where studies took place from 2007 to 2011.