Meihua Long

Experience

• 

  China Southwest Securities Co., Ltd.

  Jan 2017 - Feb 2017

  Investment Assistant

  · Through the analysis of the stock market announcement on that day, forecasted the stock price trend for further speculation in the investment advisory department;· Based on the main financial market indices, analyzed the medical device industry by using financial software Wande, and further selected individual stocks to complete the stock selection of the medical device industry.

• 

  Shining Midas Investment Management

  Jul 2017 - Aug 2017

  Data Scientist Intern

  · Utilized logistic regression and random forest to predict the rate of return on the financial instrument as a reference for portfolio managers to make investment decisions efficiently· Summarized and visualized 1000000×110 financial asset data in Python and Tableau to do data mining · Conducted dimension reduction with Kernel PCA; presented business insights to the team· Improved R square of return rate prediction by employing logistic regression with ridge and random forest model · Analyzed the user interaction experience of ValueGo (an intelligent corporate valuation client system) by using analytic hierarchy to identify areas of improvement to create better customer experience Show less

• 

  Haolan Information Technology co. LTD

  Feb 2019 - May 2019

  Machine Learning Engineer Intern

  · Built Inception neural networks to determine the infection of H. pylori bacteria, which could save each patient $200· By employing Keras increased original 804 gastroscopic images of 134 patients to 12591 images· Used TensorFlow and AWS EC2 to train deep neural networks Inception v3 and Inception ResNet v2, predict whether patients were infected with H. pylori and the degree of their infection· Visualized the training process and outcome with TensorBoard· Achieved 84.9% of best prediction accuracy for predicting whether a patient is infected with H. pylori bacteria Show less

• 

  Boeing

  Jun 2020 - Aug 2020

  Data Science & Analytics Intern (Remote)

  · Utilized Neo4j and DGL(Deep Graph Library) to build up the supply chain network (930k nodes) for Boeing’s commercial airplane sector, conducted missing links prediction using graph convolution network (GCN)· Predicted missing links based on the parts similarity match, imputed 5% of missing parts assemblies · Built Neo4j database in Docker, wrote Cypher to visualize the degree and depth distribution for the nodes · Built up a graph convolutional network with my mentor to predict missing supplier names that reached accuracy more than 90%· Diagnosed the model’s problem by looking up GitHub source code, worked with my mentor to customize the code by assigning different weight matrix to central and neighbor nodes Show less

• 

  美团

  Jan 2021 - May 2023

  Machine Learning Engineer