Experience

2022/06 - 2022/09

TikTok/ByteDance, Software Engineering Intern, Mountain View, USA

• DataAML|MLSys Team, supervised by Haibin Lin and Juncheng Gu
  • Contributed to Byted-Torch, optimized Visual Transformer, ResNet_v67, VLBert models with torch.fx and conducted unit/end-to-end tests
  • Integrated torch.fx optimization into frameworks such as cruise, haggs, and titan
  • Verified fx acceleration for titan model zoo
  • Added a feature that can correlate CUDA ops to corresponding nn.Module to the current torch.profiler in order to know which are the time-consuming ops in a model

2020/09 - 2021/08

Tsinghua University, Research Assistant, Beijing, China

• THUNLP Group, supervised by Prof. Maosong Sun
  • Reliable Knowledge Graph Construction with Probabilistic Weighted Links(2nd Authored paper with Wenhao Li, accepted by China Conference on Knowledge Graph and Semantic Computing 2021)
  • Knowledge Graph Visualization Tool Development
  • Simple Word Sense Annotation Tool Development
  • Participated in Jiuge Project Development
  • Participated in Novel Object Caption Research Project

2019/06 - 2019/08

VMware, MTS Intern, Beijing, China

• R&D Department, supervised by Kevin Song

  • Set AI related questions for new graduates’ BootCamp (with other staff engineers)

    • Sentiment analysis based on clients’ emails
    • Machine learning due date prediction system for open source projects
    • AI technologies trend analysis based on web data

  • Math Knowledge Graph Web Application (with other interns)

    • Clustered 25 classes of relations from more than 2000 sections of wikidata using KMeans
    • Manually labelled middle school math knowledge data in rdf tripples format to build prototype demo
    • Built Web App Demo with Flask backend framework and the Echarts visualization tool

Back to Main Portfolio Page View My Projects