• Scrum Master. • DevOps Engineer. • Team Foundation Server Administrator. • Work on the production and maintenance of risk management software. • Suggest improvements on current products • Pick up corporate finance. • Undertake a full range of engineering tasks in the Systems development life cycle (SDLC). • Plan and schedule projects. • Handle project management and technical documentation. • Work on Internal Cloud Architecture. • Internal System Administrator.
- Oversaw the media committees and are responsible for managing the hall's media purchasing and planning needs. - Administered Computer Committee, Video Committee, Photography Committee and Technical Team. - Standardized media committee members' knowledge. - Created Hall's new official website, activities & points portal and intranet system. - Set up the rules and structure of software engineering concept of Hall's intranet system for issue of continuity. - Set up the SOP for using media equipment. - Proposed to improve Hall's media equipment within $10,000 budget and was approved and implemented by Hall Management.
- Worked on research project on Mapping of Finite Difference Time Domain (FDTD) Algorithm on to Graphics Processing Unit (GPU). - Implemented Multi-GPU processing on the calculation. - Successfully Optimized the FDTD calculation from 7 days to 13 minutes.
- Maintained school computers. - Revamped school official website and anniversary website. - Implemented Content Management System (CMS) in the websites.
2009-01 - 2009-06
Natural Language Processing & Machine Learning Technique for Stock Trend Prediction
This is my self-initiated Final Year Research Project and had been solely done.
A lot of researchers and experts from different areas have proved that there is a high relationship between the news articles and the behavior of the stock market. In this project, the experiments which capture the impact of news articles on stock market have been successfully performed. The results have shown that there is strong information quotient in news articles which affect the stock market trend prediction.
Data crawling techniques and MongoDb database are incorporate in the application architecture. News articles are aligned to the stock trends basic on the idea of Efficient Market Hypothesis. The possible combinations of phrases are formed by Part of Speech Tagging. The selection of meaningful words and phrases is based on statistics. Finally, the relationship between the news articles and direction of the stock trends are learned through machine learning. Different type of machine learning model and different parameter are executed to find the best prediction method. In particular, a market simulation using real-life data is conducted. The study shows that there is a high relationship between news articles and the direction of the stock trends. Furthermore, by monitoring this relationship, actionable decisions could be made.