👋 I am Wyett, a new grad student looking for Full Time Employment. I can:

This is an LSTM model that predicts the short-term price of four commercial banks in China. The model arises from Yufeng Chen, Jinwang Wu, and Zhongrui Wu. The daily closing price of four banks is trained with a multi-step output static prediction method to project the daily closing price of the next five days. Currently, the model has a mean absolute error of 0.087 and a mean square error of 0.012.

Developed a MLP model that predicts the short term price movement of the S&P 500 market index. The model arises from that developed by Xiao Zhong and David Enke. 50+ data sets are pre-processed with PCA to reduce the dimensionality. Then a three-layered feedforward ANN structure with tangent sigmoid transfer function was used to output the probability of the movement of the index.

Actively running a $200,000 portfolio for my parents; since real money is put into action, I conduct extensive research and due diligence before each trade. Currently, the portfolio consists of Fixed Income and Equity Long strategies. I also use the MLP algorithm mentioned above to mitigate risk. When the MLP model predicts market downturn, a put option is bought for the underlying asset.

This is an AI hackathon where each team needs to propose a solution that tackles a particular issue identified with NLP technology. Our team decided to focus on the career development prompt, and proposed a product that provides university-specific resources to students looking to advance their careers. During the competition, we were given access to Open AI's API features and created a simple demo of how the newest transformer models may be used in our proposed solution.

This is a Data Science/AI hackathon where teams receive three data sets to analyze and study. Afterward, the teams must create a model that can predict unlabeled data. During the hackathon, our team tried models such as linear regression, logistical regression, MLP, and random forest. Random forest is used at last in the end with F1 score = 71.89.

Used C# and Unity to create a 2D indie game with the school’s team during Global Game Jam Competition.

This very website, written in HTML & CSS, and hosted on Amazon Webservice (AWS).

Used Python and Tkinter library to create the control algorithm and GUI of a self-made octa-drone and several vehicles in the maker-space club.

PS: sadly, it crashed 5 seconds after the first attempted take off.

Used Java to create an encryption and decryption application that allows users to encrypt messages with different encryption algorithms such as SHA256 or simple substitution.

Used Racket, a dialect of Lisp, to create a compiler that turns WLP4 (a simplified version of C) code into MIPS machine code.

With engineers from Doodle Mobile Ltd., using Java, created a game level algorithm that employed many techniques from nodal networks to produce high-quality levels.

Used C++ to independently create a card game “Straights” as the final project for CS246 class.

While working at Doodle Mobile Ltd., created Python web crawlers that legally download data from websites, including useful information to the company or information about our competitors’ products.