Skip to content

English Version

Work Experience

Sina Weibo ~ Beijing, China ~ 11/2014 - Present

  • Machine Learning Algorithm Engineer
    • Developed user profiling and interest prediction models
    • Improved short-term interest prediction from daily to real-time (within 10 minutes)

Yahoo! Research ~ Barcelona, Spain ~ 02/2014 - 07/2014

  • Research Intern
    • Published paper on applying space syntax to online mapping tools

Education

UPC & Université Lumière Lyon 2 ~ Barcelona, Spain & Lyon, France ~ 09/2012 - 09/2014

  • Master's in Data Mining and Knowledge Management (DMKM)
    • Erasmus Mundus Scholarship
    • Key courses: Kernel-based Learning, Statistical NLP, Advanced Statistical Modeling

Zhejiang University ~ Hangzhou, China ~ 09/2008 - 06/2012

  • Bachelor's in Mathematics and Applied Mathematics
    • GPA: 3.87/4.0, Top 5%

UCLA ~ Los Angeles, USA ~ 08/2009 - 09/2009

  • Summer School
    • Courses in Finance and English Writing

Publications

[P1] Apply Space Syntax to Online Mapping Tools, WSDM 2017

  • Yandi Li, Nicola Barbieri (Tumblr), Daniele Quercia (Bell Labs)
  • Key Technologies: Factorization Machine, BPR, PostgreSQL, PostGIS

[P2] Different Implementations and Comparisons of the Chebyshev-Tao Method

Projects

User Modeling ~ 05/2022 - Present

  • Developed Deep Learning Models for User Interest
    • Denoising VAE, Caser structure, audience normalization, unified feature embedding
  • Build Comprehensive user profiles
  • Key Technologies: Fuxi-CTR, PyTorch, Ollama, Huggingface, Spark

Weibo User Interest ~ Technical Lead ~ 05/2019 - 05/2022

  • Long and short-term interest retrieval
    • Improved real-time interest updates, integrated more behavior logs
    • Challenges: sensitivity, Matthew effect, granularity issues (Patent CN115827966A)
  • Ad targeting: transitioned from statistical models to supervised algorithms (GBDT), improving targeting conversion rates
  • Key Technologies: Flink, Hivemall, LightGBM, Grafana, Streamlit, Clickhouse

Image Recommendations ~ Technical Lead ~ 08/2017 - 04/2019

  • Implemented end-to-end image recommendation
  • Developed models for image semantic vectors using contrastive learning
  • Key Technologies: asyncio, sanic, PyTorch, Spark, faiss

Image Feature Extraction ~ 08/2016 - 10/2017

  • Developed image classification and face recognition
  • Fraud detection: identify fake celebrity avatars
  • Intelligent image cropping for thumbnails
  • Key Technologies: TensorFlow, Keras, Docker, CNN

Article Recommendation ~ 03/2015 - 11/2016

  • Built text classification(TextCNN using title, summary, body, author dimension, multi-model stacking) and clickbait detection models
  • Similar image deduplication system
  • Key Technologies: Keras, scikit-learn, Elasticsearch, MySQL, phash

Personal Projects

KDD Cup (Authorship Disambiguation) ~ Lyon, France ~ 03/2013 - 06/2013

  • Disambiguated author names in a large academic database, [Report]
    • Multilingual text processing, LDA topic extraction
  • Key Technologies: R, PostgreSQL, Python, \(\LaTeX\)

Yet Another Datalog Interpreter ~ Lyon, France ~ 09/2012 - 06/2013 - Key Technologies: Datalog, Ocaml, SQL, [Report]

Optimal Rescue Search Path ~ Hangzhou, China ~ 2010 - 2012 - Key Technologies: Graph theory, Hamilton path, [Report]

Skills and Interests

  • Programming Languages: Python, SQL, R, Matlab, Java
  • Tools and Frameworks: PyTorch, Keras, scikit-learn, Flask, ELK, Flink, Docker
  • Hobbies: Tennis🎾, Ping Pong🏓️, Tai Chi☯️, Skiing🎿, DIY🔧

Download PDF