Chao Jun LIU
Curriculum Vitae
Completed whole curriculum one year in advance.
Granted with Excellence Scholarship continuously for two years.
Granted with Excellence Scholarship continuously for two years.
The Data Science track develops strong mathematical, statistical, and computational and programming skills through the general master's core and programming requirements, in addition to providing fundamental data science education through general and focused electives requirement from courses in data sciences and related areas.
This program tackles different fields from a mathematical perspective: Big Data, Machine Learning, Statistical Modelling, Optimisation, Stochastic Process, Advanced Algorithm and a set of computer programming skills.
- Machine Learning: classical machine learning algorithms and deep learning.
- Data Mining: Feature engineering, Data visualisation, Graph mining, text mining.
- Optimisation: Convex programming, Integer programming, Stochastic programming, heuristics.
- Familiar: Python, R, scala, spark, C++, mySQL, Linux, CSI, Shiny, Latex
- Elementary: Java, mongodb, Neo4j
- English: Excellent proficiency
- Chinese: Mother Tongue
- French: Professional proficiency
- Finalist prize in Mathematical Competition in Modelling. An international competition organised by the US.
- Stochastic signal transformation. Using stochastic model and Python to simulate how signals are transmitted among nodes, optimising the capacity of wireless networks.
- Prediction of Missing Links in Citation Network. Construction of features (TF-IDF Similarity, Nnumber, etc.) based on graph and text model; Evaluation of performance of various models (Multilayer Neural Network, Random Trees Classifier, etc.)
- Handwritten Digits Recognition with Kernel Method. Implementation in Python of kernel algorithms (Kernel Ridge Regression, SVM and kernel PCA, etc.) only with NumPy and SciPy package.
- Implementation of SQL Query with MapReduce. Implementation of basic SQL query (join, sort, etc) ̇ with MapReduce in Hadoop File System in Java.
- Building a pricing tool for an international medical insurance product with expatriate contracts. Machine learning approaches are used to precisely estimate the undertaken risk head by head.
Data Scientist (Global Graduate Program)
AXA
October 2016
to October 2017
- Develop predictive models to anticipate the behaviour of customers
- Build internal risk-monitoring models, predict and analyse claims
- Perform R&D studies on the main risks of our activity as on innovative risks
- Detect events in policyholder’s lives in order to offer leading business solutions designed to meet the special needs of customers
- Ensure portfolio profitability
- Monitors portfolio risks to best identify and mitigate them
Data Scientist
AXA Global P&C
March 2016
to October 2016
Paris
France
- Deal with the curse of high dimensionality of dummy variables resulted from the high cardinality of categorical features in a pricing model.
- Create an R library in side the company in order to Implement the algorithm for sharing purpose.(private access from Gitlab).
Schedule & Dispatch Specialist
EMC
August 2013
to August 2014
Full-time
Shanghai
China
- Manages workload across the whole APJ region including Southeast Asia, India, Australia and New Zealand.
- Managed the routes and daily schedules of 35+ Field Engineers in Optimized Regions (Singapore, New Zealand and Australia).
- Provides guidance to other team members on process alignment and adherence and acts as a workflow manager within assigned team. Disseminates process changes and updates training materials as applicable.
Data Scientist
Shift Technology
March 2015
to August 2015
Internship
Paris
France
- Perform data virtualisation by multi-dimensional histogram and clustering plot in order to understand the behaviour(distribution) on raw datasets.
- Implement a various numbers of machine learning algorithms on the extracted data set for testing purpose, including SVM, CLUSEQ, HMMs, Time Series and k-NNs.
- Adapt new mathematical algorithm to comply with the special data structure(series of dates and sequences) for optimising the insurance fraud detection methods.
