About

MS in Financial Technology and Analytics @ UT Dallas

Feature Engineering: Energy Usage Predcitor In this project, a model is developed using Light Gradient Boosting Method (LGBM) for predicting energy consumption of commercial buildings given the buildings' characteristics and local weather information. A massive dataset, provided by the American Society of Heating and Air-Conditioning Engineers (ASHRAE) consisting of over 40 million records is used to build the model. The Python programming language is used. Extensive use is made of pandas, numpy, and scipy. Several novel features are introduced, allowing very accurate predictions. Code will be available soon on my github site.

Machine Learning: Credit Risk Modeling In this project, comparative models are developed using several machine learning algorithms such as Xgboost, LGBM, and Tensorflow for predicting likelihood of credit default based on customers' payment history, demographic, social, and other factors. Code will be available soon on my github site.

Data Visualization: OmniViz aims to be a general all-purpose data visualization and analysis tool. Written in Python and leveraging the tkinter and matplotlib libraries, OmniViz organizes data in an intelligent human logical manner. An adapted version of this software, ModViz, is being used at Texas Instruments for circuit simulation data analysis and vizualization. A prototype web version, OmniWeb for vizualization and analysis of stock data has been developed; I plan to enhance this incrementally.

BS in Electrical Engineering @ UT Austin

Omnisound is an award-winning 3D audio system that incorporates headtracking to generate a highly immersive soundscape. Such a system with low latency, high accuracy, low cost, and high portability is highly sought after in the 3D audio field. After researching human auditory perception, it is found that latency exceeding 60 ms, or accuracy exceeding 8.3 degrees, would break the illusion of a 3D audio scene that adjusts to head movement. Omnisound is able to meet or exceed these objectives. The design, construction, and demonstration of the system was conducted to meet the Senior Design Project requirements of the University of Texas at Austin. The project was awarded 2nd place in the Fall 2014 UT ECE Design Contest.

Low Power Op Amp design and simulation. This project is part of the UT EE338 course and describes the design and simulation of a low-power two-stage operational transconductance amplifier (OTA) using a 180nm CMOS technology and a 1.8V power supply. It is shown that with careful design, a low-frequency gain greater than 70dB can be achieved together with high phase margin (greater than 45°), fast settling time (less than 40 ns), and low power consumption (0.467 mW). The project was awarded 2nd place in a Texas Instruments sponsored Design Contest.

Buck, Boost, and SEPIC Converter design, build, and bench test. These three projects are part of the UT EE462L course. Each of these DC-DC converters as well as their gate drive circuit are built from discrete parts. The output DC and RMS voltages, power, efficiency, and output ripple are measured on the bench and compared against theoretical values.

H-Bridge Audio Amplifier design, build, and bench test. This is a three-part project, also a part of the UT EE462L course. Each of the three components of the H-Bridge system are built from discrete parts. The output signals are measured on the bench, analyzed, and compared against theoretical results.

Etherlux: Novel Internet Controlled Smart Power Outlets (IOT) This is an ongoing project aimed to create a practical and marketable "Internet of Things" (IOT) product using various skills learned in the UT Electrical Engineering program. A fully functional internet controlled power outlet bank has been constructed and is being field tested. An ARM based single-board computer (Raspberry Pi) running an Apache web server is used to control a logic block connected to power electronics circuitry. Video

Plano West Senior High

Photothermovoltaic Solar Cell Development. This is an individual initiative project conducted between 2010 and 2011 that led to the development of a high-efficiency and cost co-optimized concentrator solar cell using a combination photovoltaic cells and thermoelectric generators. The project won several national and international awards and was named a finalist in the Intel Science Talent Search competition in 2011.