CV
Download PDF version Here
Experience
NLP Researcher - Software engineering, York University
- Project in progress: Bert-API recommender
- Paper inprocceding: Automatic Unit Test Generation for MachineLearning Libraries: How Far Are We?
- Paper inprocceding: CoCoFuzzing: Testing Neural Code Models withCoverage-Guided Fuzzing.
- ISSTA2020 Paper: CoCoNuT: Combining Context-Aware Neural Translation Models using Ensemble for Program Repair.
Software Engineer – Business Intelligence, Achievers
May. 2019 – July. 2020
Team lead of awarded project: Created Named Entity Recognition (NER) system highlighting Date, Name and Event using Spacy, NLTK, Gensim and other NLP technique
Created a recommender System that recommend nominees to the user in social network by Collaborative Filtering based on the relevance between message and user.
Reduced TeraByte-level ETL data warehouse restoration time by 24 times compared with the previous version by restructuring workflow using PostgreSQL
Increased global ETL loading speed by 16% by rewrite batch loading API using Python
Created a visualization of employee community for user influence Perdition and analysis using Python, Networkx and Pyvis
Project: Machine learning based Algorithmic trading research
May. 2019 – Present
- Run a RBC mutual fund portfolio with 9.5% annual return by analysing performance data using Python
Master Research Assistant, University of Waterloo
Dec. 2017 – Sep. 2019
Increased the accuracy of Automated Program Repair model on Quixbugs benchmark by 240% by proposing novel approach using Pytorch NMT model with CRNN layer and code mutation templates
Reduced the encoder and decoder dictionary size from 50000+ to 200 tokens and training time from 100+ hours to 20 hours by creating novel NMT architecture
Created road slippage detection model with 81% accuracy using VGG16 model and OpenCV
Software Developer, AVG Technologies
May. 2016 – Aug. 2016
Full-Stack Developer, Mitel
Sep. 2015 – Dec. 2015
Full-Stack Developer, PDCI Market Access
May. 2015 – Aug. 2015
Web Developer, Aboriginal Affairs and Northern Development Canada
Apr. 2014 – Dec. 2014
Awards
- Best Hackathon Project Award at Achievers 2019: ML-based Recommender system
- 2019 Innovation Award at Achievers
- Received Full Tuition Scholarships for Uwaterloo MASc. Software Engineering program
- Received A+ on Milestone project: AR based Poke-Mon Go using OpenCV and Unity3D
- Winner of the Twitter NLP sentiment analysis contest at UOttawa for highest model accuracy out of 70 participants
- Top 4% out of 100 participants for MNIST digits recognition challenge with 98.8% model accuracy using fine-tuned K-means model
Education
University of Waterloo, Master of Computer Software Engineering
Dec 2017 – Sep 2019
University of Ottawa, Bachelor of Computer Science, Horner with Co-op
Sep 2012 – Dec 2017
Publications
Abstraction Mechanism on Neural Machine Translation Models for Automated Program Repair
Author: Moshi Wei, Lin Tan
Sep 23rd 2019
Ensemble Learning using Convolution Neural Machine Translation for Automatic Program Repair
Author: Thibaud Lutellier, Lawrence Pang, Viet Hung Pham, Moshi Wei, Lin Tan
Jun 20th 2018