Professionals
professional
Email Verified Yu Z
0.0/5 (0 Feedback)
Member since July 20, 2020

Senior Machine Learning Engineer

  • $75.00  / hr
  • US - New Jersey
  • Save

0

In-Progress 
Projects

0

Completed 
Projects

0

Cancelled 
Projects

0

In-Progress 
Service Offerings

0

Completed 
Service Offerings

0

Cancelled 
Service Offerings

* Send inquiry to this professional or invite him/her to bid for your posted need

Send Message

Experience

Senior Machine Learning Engineer

  •  EvolutionIQ
  •  Dec 2019 - Mar 2020

• Implement AutoInt(Automatic Feature Interaction Learning via Self-Attentive Neural Networks) algorithm on insurance data, improve the AUC of the current XGBoost model by 6%.
o Skills & tools: Python, Tensorflow, XGBoost

NLP Researcher

  •  Lenovo
  •  Oct 2018 - Jun 2019

• Design and developed a BERT based relation extraction model to extract the person-organization relation from sentences, got a 95% AUC score whereas the classic w2v-LSTM architecture got 85% on test set.
• Implement BERT based Bi-LSTM-CRF model, to help identify the name entities (e.g. organization, person, location, etc.) from texts. The AUC of the model reaches 99.9%.
o Skills & tools: Python, Tensorflow, Word2Vec

Senior Machine Learning Engineer

  •  ASTRI
  •  Apr 2017 - Oct 2018

• Help HSBC risk management team build a deep learning model to analyze the HKEJ and HKEX news, use CNN model to figure out potential risk of target companies.
• Stack non-text features (e.g.: frequency of target company mentioned) onto the fully connected layer of CNN, improve AUC of model by 6%.
• Help Manulife build an efficient social network comments sentiment classification system, use RNN(GRU/LSTM) model to find out negative comments or posts.
• Use data from different domain (e.g.: e-commercial comments data) to train the sentiment analysis model by using Domain Adversarial Neural Network (transfer learning method).
• Help SAE build a production issue root cause identification system by using Random Forest model with Spark mllib and python scikit-learn libraries.
o Skills & tools: Python, Tensorflow, Java, Spark, Word2vec, scikit-learn

Software Engineer

  •  Huawei
  •  Jun 2006 - Present

• Participated in NE40E (internet backbone core router) micro system development. Implemented multi-protocol transcoding framework.
• Designed and developed the contingency module (CanBus) of NE5000E (internet backbone core router). Improved system resilience and stability.
o Skills & tools: C

Education

Master of Business Analytics

  •  University of California San Diego
  •  Oct 2014 - Jun 2016

Business Analitics

Bachelor of Automatic Control

  •  Huazhong University of Science and Technology
  •  Sep 2001 - Jul 2005
error: Content is protected.